KRAS^G12D-Specific Targeting with Engineered Exosomes Reprograms the Immune Microenvironment to Enable Efficacy of Immune Checkpoint Therapy in PDAC Patients

GMP grade exosomes for siRNA delivery showed no toxicity in preclinical studies

The oncogenic Kras targeting approach we devised employed exosomes purified from the culture supernatant of bone marrow derived stromal cells from a healthy donor (Figure 1a). The purified exosomes were subjected to electroporation for the incorporation of siRNA with a sequence targeting KRAS^G12D (Figure 1a). The resulting exosomes were then processed for quality control (QC), which included size assessment by NanoSight^TM (Supplementary Figure 1A) and flow cytometry analyses for CD47 and putative exosomal marker CD63 (Supplementary Figure 1B-C). Following electroporation with siRNA, the exosomes were referred to as ‘iExoKras^G12D’. Additional validation with exosomes marker, CD9, CD81 are also provided (Supplementary Figure 1C). The iExoKras^G12D were then quantified using microBCA assay to define the exosomal protein content used for dosing definition (discussed below). The siRNA content in the iExoKras^G12D equaled that of the exosomal protein content (1:1 by weight). We previously detailed the GMP, large-scale production of iExoKras^G12D, including quality control assessment and preclinical testing showing anti-tumor effect with increased survival in PDAC GEM¹⁷. Scanning electron microscopy of iExoKras^G12D showed a distribution of exosomes (Supplementary Figure 1D) consistent with NanoSight^TM size distribution measurements. Cryogenic electron microscopy (Cryo-EM) revealed iExoKras^G12D lipid bilayer vesicles, some of which showed multiple vesiculation (Figure 1B, Supplementary Figure 1E), consistent with recent reporting¹⁸. Keeping in mind the clinical utility of GMP iExoKras^G12D, we determined the toxicology profile of GMP iExoKras^G12D in mice and non-human primates (Rhesus macaques). Three cycles of iExoKras^G12D treatment administered intravenously, 3 doses per cycle for a total of 9 doses over the course of 6 weeks, was evaluated in healthy adult mice (Figure 1C). A comprehensive toxicology profile was established, evaluating diluent for iExoKras^G12D (PlasmaLyte) and increasing iExoKras^G12D doses (1X, 2X, and 4X, Table 1). Body weight showed no significant changes with any of the dose levels (Figure 1D, Supplementary Figure 2A), and liver and kidney weights were unchanged (Supplementary Figure 2B). Although a significant % change in spleen weight was registered for 1X iExoKras^G12D, no significant changes were noted for higher doses (Supplementary Figure 2B). Comprehensive chemistry and hematology panels revealed no changes outside physiological ranges (Supplementary Figure 2C-D). When significant changes were observed when compared to the control group, those changes also remained within normal physiological range (Supplementary Figure 2C-D).

• Download figure
• Open in new tab

Figure 1. GMP iExoKras^G12D preclinical safety and biodistribution

A. Schematic representation of iExoKras^G12D GMP production. Conditioned media (CM) from mesenchymal stromal cells (MSC) cultured in GMP was centrifuged to remove debris, filtered through 0.2 um filter, and ultracentrifuged. The exosomes pellet was resuspended, subjected to electroporation to incorporate siRNA targeting KRAS^G12D using a 1:1 ratio (by weight) of exosomal protein and siRNA, and evaluated using quality control (QC) analysis for size distribution (mean size < 400 nm) and exosome surface detection of CD47 and CD63 by flow cytometry. B. Representative picture of cryo-electron microscopy of GMP iExoKras^G12D. Scale bar: 100 nm. C. Schematic representation of murine toxicology study using intravenous (i.v.) GMP iExoKras^G12D treatment (Rx) for three consecutive cycles. Each cycle represents two weeks (W1: week 1; W2: week 2, etc.) and 3 iExoKras^G12D treatment on days 1, 4, 10. D. Body weight measurement over time for the listed groups of mice, PT: pretreatment. 1X, 2X, and 4X exosomes reflect dosing corresponding to clinical trial dosing, see also Table 1, n=10 mice per group. E. Schematic representation of NHP toxicology study using intravenous (i.v.) GMP iExoKras^G12D treatment (Rx) for three consecutive cycles. Each cycle represents two weeks (W1: week 1; W2: week 2, etc.) and iExoKras^G12D treatments on the listed days, distributing 3 iExoKras^G12D treatments over the course of 2 weeks. F. Body weight measurement over time for the listed NHP. For dosing details, see Table 1. G. Schematic representation of NHP biodistribution study using intravenous (i.v.) or intraperitoneal (i.p.) GMP iExoKras^G12D that was labeled with PKH-67 or DiR dye. H-I Representative fluorescence images (H) of pancreas and liver sections from NHP following PKH-67⁺ iExoKras^G12D treatment and associated quantification (I) of the density of the ratio of PKH67 level over nuclear staining (DAPI) in the listed organs. Scale bar: 10μm. J. Normalized relative luminescence unit (RLU) in the listed organs for the detection of DiR⁺ label following i.v. or i.p. DiR⁺ iExoKras^G12D treatment. CNS: central nervous system; GI: gastrointestinal; Med. oblongata: medulla oblongata; LN: lymph node.

Toxicology studies were also evaluated in non-human primates (NHP), following a similar dosing strategy as planned for clinical testing (Figure 1E, Table 1). Three Rhesus macaques were treated with 9 doses of GMP iExoKras^G12D over the course of 6 weeks (Figure 1E, Table 1). Body weight measurements were unchanged (Figure 1F, Supplementary Figure 3A). Organ weight, chemistry and hematological studies were minimally or insignificantly altered (Supplementary Figure 3B-C). Gross examination was unremarkable, and urine analysis and a detailed anatomic pathology evaluation revealed no lesions or histopathological changes. Furthermore, we investigated the biodistribution of exogenously administered GMP iExoKras^G12D in Rhesus macaques 24 hours following intravenous (i.v.) or intraperitoneal (i.p.) injections (Figure 1G, Table 1). In this experimental design, PKH-67⁺ labeled iExoKras^G12D-infused NHP (i.v.) served as control for DiR⁺ labeled iExoKras^G12D (and in reciprocal). PKH-67⁺ exosomes were visualized in pancreas, liver, and brain tissue sections with confocal and high-resolution fluorescence microcopy, with minimal positive PKH-67⁺ exosomes noted in kidney, lung, skin, spleen, and tongue (Figure 1I-J, Supplementary Figures 4-5). Histopathological examination of pancreas, liver, and brain was unremarkable (Supplementary Figure 6A). Whole tissue bioluminescence capture of DiR⁺ signal in NHP administered with DiR⁺ labeled iExoKras^G12D i.v. vs i.p. showed positive signal in CNS tissues, gastrointestinal tract, and immune organs including lymph nodes (Figure 1J), with minimal positive foci noted in kidney, lung, skin, spleen, and tongue (Supplementary Figure 6B). While liver capture of DiR⁺ signal was greater following i.v. administration of iExoKras^G12D, omentum capture of DiR⁺ signal was prominent following i.p. administration (Figure 1J). In this experimental setting however, high background autofluorescence of certain control tissues, including occipital and parietal lobes of the cortical brain, pancreas, and prostate, prevented accurate capture of DiR⁺ signal (Supplementary Figure 6B). Quantification for KRAS^G12D siRNA by digital PCR analysis showed detection of siRNA from iExoKras^G12D in liver, pancreas, and bone marrow (Supplementary Figure 7). Levels of siRNA were higher in the liver when iExoKras^G12D were administered i.v., with the highest level of siRNA captured in the pancreas of NHP following i.v. iExoKras^G12D, albeit RNA degradation prevented comparative analysis of siRNA in one i.v. treated NHP pancreas (Supplementary Figure 7A-B). Taken together, the results support iExoKras^G12D administration into mice and Rhesus macaques is safe, with a biodistribution indicative to iExoKras^G12D accumulation in pancreas and liver.

Systemic iExoKras^G12D treatment showed no measurable toxicity in dose escalation clinical trial

Clinical testing of iExoKras^G12D was initiated with a 3+3 dose escalation (Phase Ia) and enrolled 9 evaluable patients with metastatic pancreatic cancer at three dose levels (Figure 2A). The three dose levels were equivalent to dose levels tested in toxicology studies in mice and NHP (Table 1). iExoKras^G12D dosing was defined using exosomal protein level (which contained an equivalent weight of siRNA payload [Table 1]). Patient eligibility included histologically confirmed metastatic PDAC with KRAS^G12D mutation (confirmed in tissue or blood), progression of disease on one or more line of systemic therapy in the metastatic setting (or stable disease with at least 4 months of chemotherapy with cytotoxic therapy), and Eastern Cooperative Oncology Group (ECOG) performance status of 0 and 1. Consented patients were at least 18 years of age, with negative pregnancy serum testing for women of childbearing potential and adequate contraception for males. Exclusion criteria included pregnancy, concurrent severe or uncontrolled stable angina, myocardial infarction within 6 months, symptomatic arrythmia, uncontrolled diabetes, serious active infection, central nervous system disease (with exception of treated brain metastasis). In Phase Ia, patients received 3 doses of iExoKras^G12D over the course of 2 weeks (1 cycle) for three consecutive cycles (Figure 2B). Response was assessed by imaging using Response Evaluation Criteria in Solid Tumors (RECIST) 1.1, utilizing the Quantitative Imaging Analysis Core (QIAC), a centralized platform for diagnostic imaging at the MD Anderson Cancer Center. Dose-limiting toxicity grading followed the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE, Version 4.0). Six males and three females, with a range of 49 to 75 years of age were enrolled (Figure 2C). All patients received multiple lines of prior systemic therapies and showed time from diagnosis to cancer related death ranging from 1 year (or less) to 7 years (Figure 2C). KRAS^G12D mutation status was confirmed along with additional mutations (Figure 2C). Three patients, one each at dose level 1, 2, and 3, showed stable disease after 3 cycles and were enrolled in additional iExoKras^G12D treatment, per protocol (Figure 2B-C). One patient, enrolled in the dose 3 level, showed symptomatic progression of disease after receiving just one iExoKras^G12D infusion (less than 1 cycle) and was excluded from further analyses and replaced, given that rapid progression precluded continuation of 3 consecutive cycles (Figure 2C). All but one patient that completed at least 3 cycles of iExoKras^G12D therapy, including those with initial stable disease staging on iExoKras^G12D, eventually progressed with new lesions and/or target lesions (Figure 2C). S3D1 showed stable disease after completion of 6 cycles of iExoKras^G12D and was found subsequently to have progression of disease 3 months after completing the iExoKras^G12D trial. No related adverse events were noted for any patients at the dose levels tested in the 3+3 Phase Ia trial (Figure 2C). Given lack of toxicity in all dose levels tested, a Phase Ib with accelerated dose titration was designed, with 3 additional patients enrolled at increased dose level for each patient (Figure 2A-B). To ease patient infusion schedule, the accelerated titration Phase Ib study was revised with weekly administration, for a total of 6 infusions over the course of 6 weeks (Figure 2B). Tissue biopsies were captured pre- and post-iExoKras^G12D treatment (Figure 2B). Two females and one male in an age range of 48 to 76 years and with confirmed KRAS^G12D status were treated with increasing levels of iExoKras^G12D (Figure 3A). Timeline of disease progression, prior therapy, time of biopsy for mutation analysis, and visual display of target lesion monitoring indicated iExoKras^G12D therapy represented a short window of time compared to prior treatments (Figure 3B). All three patients showed stable disease and enrolled for additional iExoKras^G12D dosing before showing progression and growth of target lesions (Figure 3A-B). Patient disease progression in this accelerated titration Phase Ib trial was between 1 (or less) to 2 years from cancer diagnosis to cancer related death (Figure 3A). Though patients demonstrated an ECOG status of 0 and 1 prior to enrollment, Phase Ib iExoKras^G12D treatment was administered to patients with 1-to-2 years disease progression from diagnosis to cancer related death (Figure 3B), and at 4-to-8-month range of time prior to cancer related death. In the Phase Ia/b clinical trial, iExoKras^G12D was well tolerated with no treatment-related adverse events reported and the maximum tolerated dose not reached (Figure 3A), thereby demonstrating a favorable safety profile. The highest dose level tested was 4.8 mg of iExoKras^G12D containing 4.8 mg of siRNA per dose, with 6 consecutive doses totaling 28.8 mg of iExoKras^G12D over the course of 12 weeks.

• Download figure
• Open in new tab

Figure 2. GMP iExoKras^G12D clinical trial design

A. Schematic representation of dose escalation design. DL1: dose level 1; DL 2: dose level 2; etc. B. Schematic representation of dosing schedule. PR/SD: partial response/stable disease; bx: biopsy C. Patient demographic, defining characteristics (including prior therapies), treatment cycle completion, response rate, total iExoKras^G12D received (expressed as the cumulative sum of exosomal proteins), and adverse events (AEs). Dosing cycles in yellow: dosing completed. Dosing cycles in grey: no treatment. C1D1: Cycle 1 Day1; C2D1: Cycle 2 Day 1; etc. M: male; F: female; SD: stable disease; PD: progressive disease; NL: new lesion; Dx -> Dt: time in year from diagnosis (Dx) to cancer related death (Dt). G+A: gemcitabine + Abraxane; XRT: radiation therapy; GemOx: gemcitabine + oxaliplatin; [CT]: other clinical trial; NAT: neoadjuvant therapy; 5FU: 5-fluorouracil; n/a: not applicable.

• Download figure
• Open in new tab

Figure 3. GMP iExoKras^G12D in accelerated titration dosing design

A. Patient demographic, defining characteristics, treatment cycle completion, response rate, total iExoKras^G12D received (expressed as the cumulative sum of exosomal proteins), and adverse events (AEs). Dosing cycles in yellow: dosing completed. Dosing cycles in grey: no treatment. M: male; F: female; SD: stable disease; PD: progressive disease; NL: new lesion; Dx ® Dt: time in year from diagnosis (Dx) to cancer related death (Dt). n/a: not applicable. B. Schematic representation of treatment timeline for patients in DL 4: dose level 4, DL 5: dose level 5, and DL5, and DL6: dose level 6, depicting prior therapy, timeframe for iExoKras^G12D therapy, and target lesion progression; scale bar listed. C1D1: Cycle 1 Day1; C2D1: Cycle 2 Day 1; etc. QW: weekly; Dx: diagnosis; mo: month; Dt: death; [CT]: clinical trial; **: mutation analysis.

Systemic iExoKras^G12D treatment showed target engagement and remodeled tumor immune microenvironment

We previously showed GMP iExoKras^G12D targeting Kras^G12D in PDAC GEM synergized with gemcitabine chemotherapy to limit tumor growth and increase survival¹⁷. We first conducted preclinical studies to determine if serial liquid biopsies for circulating Kras^G12D DNA would serve as a surrogate for tumor response. Following orthotopic cancer cell injection, mice were monitored for tumor burden by IVIS imaging. Prior to treatment start (pre-tx), mice were bled (Figure 4A). The level of circulating Kras^G12D DNA was compared in blood collected prior to iExoKras^G12D treatment and 9 days after treatment, with treatment groups including gemcitabine, GMP iExoKras^G12D, or a combination of iExoKras^G12D and gemcitabine treated mice, and with control group including mice treated with diluent or control, non-electroporated exosomes (CE: control exosomes)¹⁷. As cancer progressed in control groups, the relative level of Kras^G12DDNA (expressed as a percent of Kras^G12D copies over total Kras copies) increased with tumor burden (Figure 4B). Suppression of tumor growth, with gemcitabine, iExoKras^G12D, or combination thereof, tracked with lower relative levels of circulating Kras^G12D (Figure 4B). iExoKras^G12D therapy and combination therapy with iExoKras^G12D showed lowest relative levels of circulating Kras^G12D (Figure 4B), congruent with target specific engagement and anti-tumor response¹⁷. Mimicking the preclinical data, assessment for relative levels of circulating Kras^G12D in patients treated with iExoKras^G12D in the Phase I trial demonstrated reduced mutant allele fraction in several patients, including S2D2, S1D3, S4D3, as well as S1D5 (Figure 4C-D).

• Download figure
• Open in new tab

Figure 4. cfDNA and biopsy analyses

A. Schematic representation of murine experimental timeline for tumor inoculation, IVIS imaging of tumor burden, and blood collection. B. Relative level of Kras^G12D cfDNA (expressed as a percent of Kras^G12D copies over total Kras copies) in pre-treatment (pre-tx) and post-treatment (post-tx) with iExoKras^G12D and/or gemcitabine. Each line represents a distinct mouse per the listed experimental group: Control n=6; iExoKras^G12D n=4; gemcitabine n=6; iExoKras^G12D + gemcitabine n=7 mice. C-D. Relative level of KRAS^G12D cfDNA in the listed patients over time in the 3+3 (C) and accelerated dose titration (D) studies. PRE: pre-treatment; POST: post-treatment, with cycle definition at each time point evaluated. E. Representative H&E-stained tumor section from metastases in the abdominal wall (abdomen) or liver pre- and post-treatment with iExoKras^G12D. Scale bar upper panel: 200 μm, lower panel 100 μm. F. Representative pERK-stained tumor section from metastases in the abdominal wall of S1D4, scale bar listed.

Tissue biopsies pre- and post-treatment were captured in the accelerated titration Phase Ib trial, and sections comprising of cancer cells and stromal cells, indicated by H&E staining (Figure 4E), were selected for further immunolabeling studies. Down regulation in phosphorylated ERK, a downstream effectors indicative of active RAS signaling (Figure 4F), decrease number of PanCK⁺ cancer cells, and stable or relative increase in αSMA⁺ stromal cells noted post-treatment (Figure 5A-B). Analysis of T cell infiltrates revealed an increase in intratumoral CD8⁺ T cells, CD4⁺ Foxp3⁺ Tregs and CD4⁺ Foxp3^- cells post-treatment compared to pre-treatment (Figure 5C-D, Supplementary Fig. 8A). Taken together, these results highlight the consistent target engagement of iExoKras^G12D in human PDAC with remodeling of the tumor immune microenvironment that likely favors to anti-tumor response, similar to that observed in murine PDAC treated with iExoKras^G12D4,5.

• Download figure
• Open in new tab

Figure 5. Target engagement with iExoKras^G12D treatment

A-B. Representative images of immunolabeling for pERK, PanCK, αSMA (A), with quantification of percentage of pERK⁺ panCK⁺ cancer cells (out of panCK⁺ cells) and αSMA⁺ pixels (B) in paired biopsies. C-D. Representative images (C), and quantification (D) of CD8⁺ T cells infiltrates in paired biopsies. Data are presented as mean. Scale bar:100 μm.

CD8⁺ T cell recruitment and CD4⁺ T cells or Tregs depletion favored iExoKras^G12D therapy in preclinical models of PDAC

Informed the clinical data therein, we tested the efficacy of iExoKras^G12D therapy in combination with immune checkpoint blockade to enhance overall anti-tumor response. GMP-iExoKras^G12D downregulated KRAS^G12D transcript in Panc-1 cells (Supplementary Figure 8B). Given the increase in intratumoral CD8⁺ T effectors and CD4⁺ Foxp3⁺ Tregs in clinical trial biopsies following iExoKras^G12D therapy, we next tested the functional contribution of these cells in PDAC following Kras^G12D inhibition with iExoKras^G12D. T cell infiltration following Kras^G12D inhibition in established tumors of the autochthonous KTC (P48-Cre, LSL-Kras^G12D/+, Tgfbr2^F/F)¹⁶ PDAC GEM was measured (Figure 6A). Analysis of the intratumoral T cell populations revealed an increase in infiltration of CD4⁺ and CD8⁺ T cells in the PDAC tumor microenvironment (TME) following iExoKras^G12D treatment (Figure 6B-C), consistent with the findings in clinical samples (Figure 5E-I). Similar results were also noted in a second orthotopic KPC-689 PDAC model treated with iExoKras^{G12D 16} (Supplementary Figure 9A), resulting in increased tumor infiltrating CD4⁺ and CD8⁺ T cells with iExoKras^G12D treatment (Supplementary Figure 9B). To determine the functional contribution of the T cell infiltrates following Kras^G12D inhibition, we crossed the KPC (Pdx1-Cre, LSL-Kras^G12D/+, Trp53^R172H/+) GEM with CD4^−/− or CD8^−/− mice and generated KPC mice depleted of CD4⁺ T cells or CD8⁺ T cells, respectively. KPC, KPC CD4^−/− and KPC CD8^−/− mice were treated with iExoKras^G12D and tumor burden was monitored by serial MRI (Supplementary Figure 9C-D). The control (no iExoKrasG12D) KPC, KPC CD4^−/− and KPC CD8^−/− mice showed similar tumor progression and survival kinetics (Supplementary Figure 9D-E). KPC mice treated with iExoKras^G12D demonstrated prolonged survival compared to control KPC mice (Supplementary Figure 9E), consistent with our prior findings¹⁶. KPC CD4^−/− mice treated with iExoKras^G12D demonstrated robust tumor growth inhibition by MRI and prolonged survival compared to control KPC CD4^−/− mice and iExoKras^G12D treated KPC mice (Supplementary Figure 9D-E). In contrast, KPC CD8^−/− iExoKras^G12D treated mice showed no survival benefit compared to the control KPC CD8^−/− mice (Supplementary Figure 9D-E), indicating that CD8⁺ T cells are required for Kras^G12D inhibition mediated suppression of PDAC⁵. Taken together, our data demonstrate that depletion of CD4⁺ T cells/T regs prime Kras^G12D inhibition and effector CD8⁺ T cells are required for PDAC control with iExoKras^G12D therapy.

• Download figure
• Open in new tab

Figure 6. Synergy of iExoKras^G12D with anti-CTLA-4 for tumor suppression in preclinical studies

A. Schematic of experimental timeline using KTC GEM. iExoKras^G12D treatment was initiated at day 33 of age until day 44. B-C. Representative images (B) and respective quantification (C) of CD4⁺ T cells (CD4), and CD8⁺ T cells (CD8) by immunolabeling in control and iExoKras^G12D treated KTC mice, n=3 mice per group. D-E. Representative images (D) and respective quantification (E) of CK19⁺ and Fas⁺ cells by immunolabeling in control and iExoKras^G12D treated mice, n=3 mice per group. F. Schematic of experimental timeline using orthotopic PKC mice. Treatment was initiated at day 12 post tumor implantation and mice euthanized 2 weeks post-treatment (day 26). G. Tumor burden ((Pancreas or tumor weights /body weight) *100) of PKC mice treated with iExoKras^G12D, control exosomes (CE), Isotypes control (Iso), αCTLA-4 and αPD-1, n=5 mice per group. H-I. Representative images (H), and quantification (I) of CK19 and pERK immunostainings in the indicated groups, n=3-5 mice per group. J. Representative image and respective quantification of CD8, CD19, CD31, and nuclear staining (DAPI) in tertiary lymphoid structure (TLS) in the tumors of mice in the indicated groups, n=3-5 mice per group. Scale bar: 100 μm. Data are presented as mean +/– SD. Significance was determined by unpaired t-test (B, C) or one-way ANOVA with Dunnett’s multiple comparisons test (E, G). αCTLA-4, anti-CTLA-4; αPD-1, anti-PD-1.

iExoKras^G12D efficacy relies on Kras^G12D primed epigenetic regulation of Fas for Fas-FasL anti-tumor response by CD8⁺ T cells

Fas expression in pancreatic cancer cells following Kras^G12Dgenetic suppression enables interaction with FasL-expressing CD8⁺ T cells and anti-tumor response^4,5. CK19⁺ pancreatic cancer cells showed an increase in FAS expression (FAS⁺ CK19⁺) following iExoKras^G12D treatment in KTC GEM (Figure 6D-E). In vitro, KPC-689 cells treated with iExoKras^G12D demonstrated upregulation of expression in Fas and the CD8⁺ T cell recruiter cytokine, IL-15 (Supplementary Figure 9F). Mechanistically, the regulation of Fas by Kras^G12D inhibition in pancreatic cancer cells included epigenetic remodeling, implicating methylation of the histones and recruitment of transcriptional repressor at the promoter site^4,19,20. Chromatin immunoprecipitation (ChIP analysis) at the Fas promoter or transcriptional start site revealed that Kras^G12D expression recruited DNA methyl transferase DNMT1 and histone methyl transferase EZH2, and iExoKras^G12D treatment inhibited DNMT1 and EZH2 recruitment (Supplementary Figure 9G). EZH2 promotes trimethylation of histone H3 at the K27 site²¹ and iExoKras^G12D treatment downregulated H3K27me3 and upregulated acetylation of H3K27 within the Fas promoter (Supplementary Figure 9G). Collectively, our data indicate that iExoKras^G12D treatment relieved Kras^G12D mediated epigenetic suppression of Fas in pancreatic cancer cells, enabling permissive immune TME priming to anti-tumor response via CD8⁺ T cell FASL-FAS mediated eradication of cancer cells.

Clinical trial with iExoKras^G12D informed on combination therapy with immune checkpoint blockade

Based on the data collected, we next tested the efficacy of iExoKras^G12D therapy in combination with ICB to enhance anti-tumor response. Mice with orthotopic PDAC tumors (utilizing syngeneic cells derived from P48Cre, LSL-Kras^G12D/+, Trp53^F/+ GEM) were treated with iExoKras^G12D and ICB (Figure 6F). Specifically, iExoKras^G12D treatment was combined with either anti-CTLA-4 or anti-PD-1 checkpoint immunotherapies (Figure 6F). Age matched analysis of PDAC tissues revealed a decrease in tumor burden with iExoKras^G12D therapy, with further decrease in tumor burden following anti-CTLA-4 combination but not with anti-PD-1 combination (Figure 6G). Immunolabeling of tumor sections revealed downregulation of pERK in all groups that received iExoKras^G12D (Figure 6H-I). Analysis of CK19⁺ cells (used here as surrogate for quantification of pancreatic cancer cells) revealed a decrease in CK19⁺ cancer cells following iExoKras^G12D, with further decrease in mice treated with anti-CTLA-4 and iExoKras^G12D combination (Figure 6H-I). Histological analysis of PDAC tissues demonstrated an increase in the presence of tertiary lymphoid structures (TLS) in mice treated with iExoKras^G12D (iExo) + anti-CTLA-4 with 100% penetrance (Supplementary Figure 10A-B). The presence of TLS, characterized by formation of new lymphoid structures including B cells, CD8⁺ T cells, a newly developing vasculature with endothelial venules and antigen presenting cells^22–24, is frequently associated with favorable responses to immunotherapy in the context of cancers such as melanoma, non-small cell lung carcinomas and better prognosis^23,24. We performed TSA multiplex immunostaining analysis confirming the presence of CD8⁺ T cells, B cells, and endothelial cells in TLS accumulating in tumors of mice treated with iExoKras^G12D + anti-CTLA-4 (Figure 6J). Collectively these results support the superior anti-tumor response of iExoKras^G12D when combined specifically with anti-CTLA-4 therapy.

Exosomes source and production

GMP iExoKras^G12D were derived from the culture supernatant of human bone marrow mesenchymal stromal cells. The procedures to produce GMP-grade iExoKras^G12D were detailed previously¹⁷. Briefly, conditioned media from human bone marrow mesenchymal stromal cells grown in a bioreactor was subjected to filtration and centrifugation. The exosomes were characterized by NanoSight^TM NS300 (Malvern) and flow cytometry analyses, and subsequently electroporated with siRNA¹⁷. MicroBCA analysis was used to define GMP iExoKras^G12D doses, expressed as mass of exosomal protein. Flow cytometry analyses of GMP iExoKras^G12D was also detailed in prior report, and gating strategy is shown in Supplementary Figure 1B.

Clinical study design and patient population

The iEXPLORE (iExoKras^G12D in Pancreatic Cancer, NCT03608631) clinical trial is a first-in-human, open-label, single-center Phase I study of mesenchymal stromal cells-derived exosomes with KRAS^G12D siRNA evaluating the safety, tolerability, and target engagement in patients with metastatic pancreatic ductal adenocarcinoma harboring the KRAS^G12D mutation. The study was conducted according to the principles of Good Clinical Practice with prior approvals obtained from the Institutional Review Board (IRB) of MD Anderson’s Human Research Protection Program (HRPP) and with investigational new drug (IND) authorization granted by the Food and Drug Administration (FDA).

The study enrolled patients with pre-defined criteria, including 18 years of age on the day of consenting to the study, and histologically confirmed metastatic PDAC harboring KRAS^G12D mutation, as informed from any previous routine molecular profiling (e.g. Foundation One) of tissue or blood. Patients had documented progression of disease on one or more lines of systemic therapy. If with stable disease, patient must have had completed at least 4 months of chemotherapy with cytotoxic therapy. Patients presented with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 and adequate organ functions. Women of childbearing potential (WOCBP, defined as not post-menopausal for 12 months or no previous surgical sterilization) had a negative serum pregnancy test within one week prior to initiation of treatment and requested to be using an adequate method of contraception to avoid pregnancy throughout the study and for up to 12 weeks after the last dose of study drug to minimize the risk of pregnancy. Exclusion criteria included concurrent severe and/or uncontrolled medical conditions that could compromise participation in the study, such as unstable angina, myocardial infarction within 6 months, unstable symptomatic arrhythmia, uncontrolled diabetes, serious active or uncontrolled infection, pregnancy or lactation, known CNS disease (except for treated brain metastasis, unless the patient was treated by neurosurgical resection or brain biopsy taken within 3 months prior to Day 1). Patient demographic data are detailed in Figure 2C and Figure 3A.

The clinical product starting dose of 0.150 mg of exosomal protein (which included an equivalent 0.150 mg of siRNA) was informed by non-toxic dosing in preclinical studies and applying conversion to human equivalent dosing (Table 1). iExoKras^G12D were generated by the Stem Cell Transplantation and Cellular Therapy Laboratory at MD Anderson according to Good Manufacturing Practice (GMP) and clinical grade siRNA was purchased from City of Hope RNA synthesis core. The iExoKras^G12D diluent was PlasmaLyte. Infusion rate of iExoKras^G12D was set at 2 mL/min, with infusion time ranging from 5min to 64min for increasing dose levels. Each iExoKras^G12D infusion was followed by PlasmaLyte infusion (100 mL over the course of approximately 30min). The initial Phase Ia study enrolled 3 patients at the dose levels (Dose Levels 1, 2, and 3) listed in Figure 2A. Given that no toxicity was observed, the Phase Ib study enrolled patients in an accelerated titration design, with a single patient at each of the increasing dose levels (Dose Levels 4, 5, 6). Phase Ib enrollment included capture of pre- and post-treatment tissue biopsies (Figure 2B). Toxicity scoring followed the Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0 for toxicity and adverse event reporting. Dose limiting toxicity (DLT) were defined as grade 3 hematologic toxicity with bleeding, grade 4 thrombocytopenia, grade 4 neutropenia of ≥7 days duration or ≥ grade 3 neutropenia of any duration with fever ≥ 38.5°C, grade 2 or greater autoimmune reaction, grade 2 or higher hypersensitivity reaction, grade 4 or higher infusion-related reaction, grade 3 or 4 events, excluding alopecia, nausea, vomiting, and/or diarrhea, unless these occur despite maximal prophylaxis and/or treatment. No DLT were encountered and MTD was not reached.

Clinical response was measured with CT imaging according to RECIST 1.1 criteria, approximately at every 3 cycles, as shown in Figure 2B. Target lesions and non-target lesions were ascertained, with Partial Response (PR) defined as at least a 30% decrease in the sum of diameters of target lesions (taking as reference the baseline sum diameters); Progressive Disease (PD) was defined by at least a 20% increase, and Stable Disease (SD) defined as neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD. Appearance of new lesion(s) (NL) also defined PD.

Non-human primate (NHP) preclinical studies

All procedures were reviewed and approved by the Institutional Animal Care and use Committee (IACUC) at MD Anderson Cancer Center (MDACC) and NHPs were housed in AAALAC-accredited facility at Michale E. Keeling Center for Comparative Medicine and Research (KCCMR) at Bastrop, TX. Two preclinical studies are reported, using GMP iExoKras^G12D, and prospectively written protocols were followed. Experiment 1 studied the toxicology of GMP iExoKras^G12D (Figure 1E). Experiment 2 studied the biodistribution of GMP iExoKras^G12D (Figure 1G).

Experiment 1: GMP-iExoKras^G12D was administered to adult female and male Rhesus macaques (Macaca mulatta) of Indian origin (4.6 to 6.1 kg body weight, 4-5 years old). Rhesus macaques were transferred from the KCCMR Rhesus Macaque Breeding and Research Resource Colony (RMBRR) to study housing acclimation prior to initiation of the pre-clinical study and were randomly assigned to 2 groups. NHP (13-069) in group 1 received PlasmaLyte (diluent, control), and NHPs in group 2 received 0.15 mg GMP iExoKras^G12D. All doses were administered in 2 ml volume, intravenously at a peripheral venous access point, and flushed with 0.5 ml volume of PlasmaLyte. A total of 9 doses were administered over the course of 42 days (see Figure 1E). The NHPs were euthanized one week after the final dose. Body weight (longitudinal and at experimental endpoint), organ weights, chemistry and hematology panels were completed. Gross necropsy findings and histopathological examination of tissues were performed following the proceeding standardized by the KCCMR.

Experiment 2: DiR and PKH-67 labeled GMP-iExoKras^G12D were administered to adult (approximately 7 years of age) male Rhesus macaques (Macaca mulatta) of Indian origin (∼6 kg body weight). For DiR labeling of GMP-iExoKras^G12D, exosomes were diluted to a final concentration of 5 billion exosomes per 1 ml of PBS and 1 µl of XenoLight DiR (Perkin Elmer, mg/mL) added. The samples were incubated for 15min at 37°C followed by incubation at 4°C for 5min. Labeled exosome samples were then diluted in 10 ml of PBS and centrifuged in a SW41Ti rotor (Beckman Coulter) at 200,000 x g for 3h at 4°C. PKH-67 labeling was performed according to manufacturer’s instructions (Sigma Aldrich). Briefly, exosomes were diluted to a final concentration of 5 billion exosomes per 1 ml in diluent C and 4 µl of PKH-67 added. Exosomes were incubated for 5min at room temperature followed by addition of 2 ml of 1% BSA in water. Samples were then diluted in 8 ml PBS and centrifuged in a SW41Ti rotor at 200,000 x g for 3h at 4°C. Following centrifugation, the supernatant was removed, the exosome pellet resuspended in PlasmaLyte, and exosome concentration evaluated by Nanosight^TM NS300 for both DiR and PKH labeled samples.

Rhesus macaques were transferred from the KCCMR RMBRR to study housing acclimation prior to initiation of the pre-clinical study and were randomly assigned to 3 groups. NHP (13-048) in group 1 received GMP iExoKras^G12D (siRNA Kras^G12D) labeled with PKH-67, intravenously, NHP (13-052) in group 2 received GMP iExoKras^G12D (siRNA Kras^G12D) labeled with DiR, intravenously, and NHP (13-072) in group 3 received GMP iExoKras^G12D (siRNA Kras^G12D) labeled with DiR, intraperitoneally. All doses for the biodistribution studies contained approximately 70 mg of exosomal protein (or approximately 70 billion exosomes) administered in 2.5 mL volume (Table 1). Rhesus macaques were euthanized 24h post dose with urine, blood, and organs collected. Organs were processed for formalin fixing and paraffin embedding, snap frozen tissues, OCT embedded tissues, and fresh tissue for IVIS imaging. For H&E analysis of paraffin tissues, 5 µm sections were cut and stained for H&E using ST Infinity H&E staining system (Leica) and Leica Autostainer XL according to manufacturer’s instructions.

NHP tissue analyses

For tissue analyses of DiR⁺ GMP iExoKras^G12D biodistribution in NHP, approximately 0.2 to 1 cm³ NHP tissue pieces were evaluated for DiR⁺ GMP iExoKras^G12D accumulation using In Vivo Imaging System (IVIS Spectrum, Perkin Elmer). Tissues were imaged 24h post injection using a 710 nm excitation filter and 780 nm emission filter with background fluorescence removed by subtracting the background measurements at time of imaging. Tissues were compared to the negative control (NHP 13-048, injected with PKH-67 labeled exosomes).

For microscopic analyses of PKH-67⁺ GMP iExoKras^G12D biodistribution in NHP, frozen sections (5 µm) were cut from tissues embedded in optimal cutting temperature (OCT) medium. Sections were stained with 0.25 µg/ml DAPI in TBS for 5min then mounted with coverslips using Fluoroshield (Sigma). Slides were imaged on a Keyence BZ-X710 fluorescent microscope and images stitched together in BZ-X Analysis software (Keyence) to generate a whole tissue image. Threshold values for green fluorescence were established based on negative control slides (NHP 13-052, injected with DiR-labeled exosomes) for each tissue and applied to stitched images in ImageJ. After applying the established threshold values to images, green fluorescence integrated density was quantified within tissue area determined by DAPI positive area. The integrated density of control slides was subtracted to calculate a normalized integrated density for each tissue.

For high resolution imaging of PKH-67 labeled exosomes, DAPI labeled sections (as described above) were imaged with a Zeiss LSM800 confocal microscope equipped with a 63x Plan-Apochromat objective. For the images presented in Supplementary Figure 5B, a z-stack spanning the entire tissue thickness at 1 µm intervals was acquired at 20x magnification and the images presented as a maximum intensity projection of the entire z-stack.

Murine preclinical studies

All procedures were reviewed and approved by the Institutional Animal Care and use Committee (IACUC) at MD Anderson Cancer Center (MDACC) and mice were housed in AAALAC-accredited facility at MDACC. All mice received LabDiet 5053 ad libitum and were housed at 21-23°C, 40-60% humidity and 12h light/dark cycle.

Two preclinical studies were reported. The toxicology study (Figure 1A) employed GMP iExoKras^G12D administered intravenously to adult wild-type C57BL/6 albino females (age 25-30 weeks). Mice were purchased from Jackson Laboratories and allowed to acclimate for 2 weeks or more prior to initiation of the preclinical study. Ten mice were randomly assigned to 4 groups each. Mice in group 1 received PlasmaLyte (diluent, control), and mice in group 2 to 4 received increasing dose of GMP iExoKras^G12D. Group 2 was administered 0.66 µg of iExoKras^G12D per dose, Group 3 was administered 1.32 µg of iExoKras^G12D per dose, and Group 4 was administered 2.65 µg of iExoKras^G12D per dose. All doses were administered in 50 µl volume, intravenously via the retro-orbital plexus. A total of 9 doses were administered over the course of 6 weeks (see Figure 1C). The mice were euthanized within one week after the final dose. Body weight (longitudinal and at experimental endpoint), organ weights, chemistry and hematology panels were completed. Gross necropsy findings and histopathological examination of tissues were performed following the proceeding standardized by the Department of Veterinary Medicine and Surgery, Division of Veterinary and Comparative Pathology at MDACC.

P48-Cre, LSL-Kras^G12D/+, Tgfbr2^F/F (KTC) GEM tumor tissue sections used for analysis in Figure 6A-C were obtained from a previously reported study¹⁶. For studies with T cell depletion mice, we crossed KPC (Pdx1-Cre, LSL-Kras^G12D/+, Trp53^R172H/+)²⁷ mice with CD4^−/− (Cd4^tm1Mak)²⁸ depleted of CD4⁺ T cells or CD8⁺ T cells CD8^−/− (Cd8^tm1Mak)²⁹. The combination preclinical study using GMP iExoKras^G12D and ICB antibodies utilized orthotopic PKC-Hy19636 (derived P48Cre, LSL-Kras^G12D/+, Trp53^F/+ mice)³⁰ cells that were injected in the tail of the pancreas of C57BL6/J mice (0.5 x10⁶ cells in 20 ml PBS using 27-gauge Hamilton syringe, Figure 6D). Both male and female adult mice were used and purchased from Jackson laboratories. For checkpoint blockade immunotherapy experiments, anti-CTLA-4 (BioXcell, Clone 9H10) or anti-PD-1 (BioXcell, Clone 29F1.A12) were injected three times per week (first dose 200 mg, followed by 2 doses of 100 mg intraperitoneal injections) in a final volume of 100 ml diluted in PBS. The control mice received a cocktail of isotype antibodies (Rat IgG2a (BioXcell, Clone 2A3) and Syrian hamster IgG (BioXcell, BE0087)) and control exosomes (non-electroporated with siRNA) in the same dosage, route and frequency as the treatment groups. For iExoKras^G12D treatments, 10⁹ GMP iExoKras^G12D (equivalent to approximately 1 µg) were injected everyday intraperitoneally in the orthotopic PKC mice and every other day in the KTC GEM study. For assessments of tumor volumes, mice were imaged at indicated time points using Bruker 7T MRI at the MD Anderson Small Animal Imaging Facility (SAIF). The mice were sacrificed at indicated time points for age-matched analysis.

Immunolabeling studies

For samples from the clinical trial, serial sections were cut through the block to collect 5 μm formalin fixed paraffin embedded (FFPE) sections. H&E analysis was performed on slides every 100 μm to select serial sections with optimal tumor tissue for immunostaining analysis. Immunohistochemistry for phosphorylated ERK (pERK, Figure 4F) were carried on 5 μm thick FFPE sections using an automated protocol (Leica Bond RX; Leica Biosystems Nussloch GmbH). The tissue sections were deparaffinized and rehydrated following the Leica Bond protocol. Antigen retrieval was performed for 20min with Bond Solution #1 (Leica Biosystems, equivalent to citrate buffer, pH 6.0). The primary antibody [Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (D13.14.4E) XP® Rabbit monoclonal antibody (Cell signaling, Cat#4370), dilution 1:400, Cell signaling, Cat#4370] was incubated for 15min at room temperature. The primary antibody was detected using the Bond Polymer Refine Detection kit (Leica Biosystems) with DAB as chromogen. The slides were counterstained with hematoxylin, dehydrated, and coverslipped.

Immunofluorescence analysis was performed on 5 μm thick FFPE sections using Tyramide signal amplification (TSA) after deparaffinization and hydration, following which antigen retrieval was performed for 20min at 98°C in Tris-EDTA buffer (pH-8.6) in a steamer. Next, the sections were incubated in 3% H₂O₂ for 10min at room temperature, 1.5% bovine serum albumin (BSA) for 1h and primary antibody at indicated concentrations for 3h (Supplementary Table 1). Following addition of the indicated secondary polymer (30min at RT), opal reagents were used at a 1:100 concentration (30min at RT) for TSA staining. Subsequently, antigen retrieval was performed before use of the next primary antibody. The slides were mounted using Fluoroshield^TM with DAPI (Sigma Aldrich F6057). For single stained IHCs, DAB reagent (5 to 15min) was added following incubation with the secondary polymer, counterstained with hematoxylin, dehydrated and mounted for analysis. For quantification of immune infiltrates, representative images were taken from 3-5 high power fields (20x magnification) per biological replicate to count the number of immune cells and average of multiple images is plotted as a replicate. We used Zeiss Axio Imager Z.2 for IF imaging. For quantification of CK19 and pERK immunostainings, average pixels positivity of 5 high power field images were used for quantification using Leica microsystems DFC295 microscope for brightfield imaging.

Digital PCR analyses for circulating Kras^G12D DNA

We previously reported on the anti-tumor efficacy of GMP iExoKras^G12D with and without gemcitabine in an orthotopic model of pancreatic cancer¹⁷. The blood of mice from the previously published experiment (Reference¹⁷, Figure 6E & Supplementary Figure 8C) was analyzed for the detection of Kras^G12D mutation in circulating DNA. Briefly, blood samples were collected from tumor bearing mice (KPC689 orthotopic tumors with Kras^G12D mutation) one day prior to the start of treatment (pre-treatment/pre-tx) and 9 days after treatment was started (post-treatment/post-tx) (see Figure 4A). Murine blood samples were collected and left at room temperature for 4h followed by centrifugation at 3,000 x g for 5min to collect serum. cfDNA from the serum for dPCR were extracted by using QIAamp DNA Micro Kit and DNA concentration was measured by Qubit dsDNA HS kit according to manufacturer’s instructions. The probe assay ID used was AH6R5PI (KRAS G12D c35G>A (WT:C —> MUT:T). 6.5 µl of DNA was added regardless of DNA concentration.

Approximately 1.5 to 4 ml plasma samples for patients on the clinical trial were also analyzed. ctDNA isolated from 0.5 ml of patient plasma using the cfPure V2 Cell Free DNA Extraction Kit (Biochain) according to manufacturer’s instructions. TaqMan Liquid Biopsy dPCR Assay (Invitrogen A44177, Assay ID Hs000000051_rm) in conjunction with Quant Studio 3D dPCR 20K was used to amplify 1 or 2 ng of ctDNA and quantify (averaging values for 1 and 2 ng) KRAS^G12D and KRAS (wildtype/unmutated allele) in ctDNA.

The analysis of the digital PCR data was performed with the manufacturer’s software (QuantStudio 3D Analysis Suite). The setting of detection limit used was the same as previously reported³¹. The results were expressed as a percent of the ratio of Target/Total (Kras^G12D/(Kras^G12D+Kras^WT)).

Quantitative RT-PCR

For in vitro experiments to determine Kras^G12D knockdown efficacy of iExoKras^G12D, 10⁵ Panc-1 or KPC-689 cells were seeded in 6 well plates (ThermoFisher, 140685) in 10% FBS DMEM with PS in a final volume of 2 ml per well overnight. Subsequently, 10⁹ iExoKras^G12D were added in serum free DMEM in a final volume of 1 ml and RNA was extracted from the cell lines using PureLink^TM RNA Mini Kit (Invitrogen, 12183025) as per the manufacturer’s instructions at indicated time points for downstream analysis. RNA concentrations were determined using spectrophotometer (Nanodrop 2000) and 500 μg to1 mg RNA was used for reverse transcription using High-Capacity cDNA Reverse Transcription Kit (ThermoFisher Scientific, 4368814). The cDNA was further diluted 1:20 for qPCR analysis with SYBR^TM green master mix (Applied Biosystems^TM, A25742). Primers used for qPCRs include:

_KrasG12D

F: 5’-ACTTGTGGTGGTTGGAGCAGA-3’ R: 5’-TAGGGTCATACTCATCCACAA-3’

Fas:

F: 5’-TATCAAGGAGGCCCATTTTGC-3’ R:5’-TGTTTCCACTTCTAAACCATGCT-3’ Il15:

F: 5’-ACATCCATCTCGTGCTACTTGT-3’ R: 5’-GCCTCTGTTTTAGGGAGACCT-3’ 18s:

F: 5’-GTAACCCGTTGAACCCCATT-3’

R: 5’-CCATCCAATCGGTAGTAGCG-3’

Gapdh:

F: 5’-AGGTCGGTGTGAACGGATTT-3’

R: 5’-TGTAGACCATGTAGTTGAGGTCA-3’

Gene expression for listed genes was normalized to 18s or Gapdh and fold change was determined by ratios of 2^-dCt values. Statistical comparisons were performed on dCt values.

Chromatin immunoprecipitation

For ChIP experiments, 0.5 to 1 x 10⁶ cells per immunoprecipitation (IP) condition from iExo and ScrExo [2 doses (10⁹ exosomes/dosage) administered 24h apart] of KPC-689 cells were collected on day 3 by trypsinization. Cross-linking was performed by shaking cells in a suspension of 1% methanol free formaldehyde (ThermoFischer, Cat# 28908) in complete medium at room temperature for 10 to 12min. Subsequently, the solution was mixed with 125 mM glycine and incubated at room temperature for 5min. The cells were washed twice with ice cold PBS (with 1 mM PMSF) and the resulting cell pellets were stored at –80^°C for downstream analysis. ChIP assays were performed at the MD Anderson’s Epigenomics Profiling Core with some modifications to previously described high-throughput ChIP protocol³². For histone ChIPs, H3K27ac, Abcam – Cat# ab4729; H3K27me3, Diagenode – Cat# C15410195 and H3, Abcam – Cat# ab1791 were used. For ChIPs for non-histone antibodies (DNMT1, Novus Biologicals, Cat# NB100-56519 and EZH2, Cell Signaling Technology, Cat# 5246S) were used. Subsequently, qPCR assays were performed using primers for the promoter/transcription start site using the respective primers (FAS TSS (Histone ChIP)-F: 5’-CTGCCTCTGGTAA GCTTTGG-3’ R:5’-CAGCCACATCTGGAATCTCA-3’; FAS promoter site (DNMT1, EZH2 ChIP)-F: 5’-CCC TGTATTCCCATTCATCG-3’ and R: 5’-ACTAGGGG AGGGGACAGAAA-3’). Normalized transcript levels (to IgG or histone H3) were used for statistical comparisons.

Electron microscopy

Scanning electron microscopy – GMP iExoKras^G12D was submitted to the High Resolution Electron Microscopy Facility for processing and imaging. Fixed samples were washed in 0.1 M sodium cacodylate buffer and treated with 0.1% Millipore-filtered cacodylate buffered tannic acid, postfixed with 1% buffered osmium, and stained en bloc with 1% Millipore-filtered uranyl acetate. The samples were dehydrated in increasing concentrations of ethanol, infiltrated, and embedded in LX-112 medium. The samples were polymerized in a 60°C oven for approximately 3 days. Ultrathin sections were cut in a Leica Ultracut microtome (Leica, Deerfield, IL), stained with uranyl acetate and lead citrate in a Leica EM Stainer, and examined in a JEM 1010 transmission electron microscope (JEOL, USA, Inc., Peabody, MA) at an accelerating voltage of 80 kV. Digital images were obtained using AMT Imaging System (Advanced Microscopy Techniques Corp, Danvers, MA). All pictures were captured with the help of the facility.

Cryo-electron microscopy – GMP iExoKras^G12D was submitted to the CryoEM Core at Baylor College of Medicine for processing and imaging. The surfaces of QUANTIFOIL holey carbon grids were plasma cleaned for 20 seconds inside an Ernest Fullam Glow Discharge. The grids were transferred to a VitroBot Mark IV plunge freezer where 2.5 µl of the iExoKras^G12D sample and 1 µl of AURION 10 nm gold tracers were applied to the carbon-side of the grid before immediately being plunged into liquid nitrogen. The vitrified sample was transferred to a JEOL 2200FS operating at 200 kV equipped with a Direct Electron DE-64 detector. Micrographs were collected at 8k x 8k resolution with a final A/pixel of 2.18. The approximate total electron dose per micrograph was 20 e-/A^2. The micrograph movies were motion corrected and gain-normalized using EMAN2³³.

Supplementary Figure Legend:

Supplementary Figure 1. GMP iExoKras^G12D characterization

A. GMP iExoKras^G12D particle size and concentration by NanoSight^TM. B. Gating strategy for GM iExoKras^G12D bound to beads. C. Histogram of flow cytometry analysis of GMP iExoKras^G12D for the indicated markers. D. Representative SEM of GMP iExoKras^G12D, scale bars listed. E. Representative cryoEM of GMP iExoKras^G12D. Scale bar: 100 nm or otherwise listed.

Supplementary Figure 2. Safety profile of systemic GMP iExoKras^G12D in mice

A-B. Percent change in body weight (A) and listed organs (B) for the listed groups of mice, PT: pretreatment. 1X, 2X, and 4X exosomes reflect dosing corresponding to clinical trial dosing, see also Table 1, n=10 mice per group. C-D. Chemistry and hematology tabulated results (C) or graphical representation (D) for the listed groups. Data presented as scatter plot about the mean. One-way ANOVA with Tukey’s multiple comparisons test (A-B) or Dunnett’s multiple comparisons test (C-D). *P<0.05; **P<0.01; ns, not significant.

Supplementary Figure 3. Safety profile of systemic GMP iExoKras^G12D in Rhesus macaques A-B. Normalized body weight over time (A) and organ/body weight (BW) ratio of the listed organs for each Rhesus macaque. C. Graphical representation of hematology and chemistry analysis of the listed Rhesus macaque.

Supplementary Figure 4. Labeled exosomes biodistribution in Rhesus macaque Representative images of pancreas, liver, and brain tissue section from Rhesus macaque with nuclear staining (DAPI) and PKH labeled exosomes with digital zoom inset.

Supplementary Figure 5. High magnification images of labeled exosomes biodistribution in Rhesus macaque

A. Representative high magnification images of the listed organs in Rhesus macaque with nuclear staining (DAPI) and PKH labeled exosomes. The panel displays negative control (unlabeled exosomes) next to PKH⁺ label exosomes-treated Rhesus macaque; this panel is also displayed in Figure 1J. B. Representative confocal microscopy images of the listed organs from Rhesus macaque.

Supplementary Figure 6. Biodistribution of exosomes in Rhesus macaque

A. Representative H&E stained images of the listed organs of each Rhesus macaque. B. Normalized relative luminescence unit (RLU) in the listed organs for the detection of DiR⁺ label following i.v. or i.p. DiR⁺ iExoKras^G12D treatment (relative to control organ without exposure to DiR⁺ iExoKras^G12D). 1: unreliability of data due to high RLU in control organ; 2: unreliability of data due to lack of background RLU detection in control organ.

Supplementary Figure 7. Circulating siRNA capture by dPCR in Rhesus macaque

A-B. Histogram representation (A) and dPCR scatter plot (B) of the percent circulating siRNA dPCR capture relative to total RNA in Rhesus macaque 24h following i.v. or i.p. administration of iExoKras^G12D. siRNA concentration curve for positive control and range of detection used 27 ng, 270 pg, 27 pg of Kras^G12D siRNA and 270 pg of scrambled (scrbl) control siRNA. Rhesus macaque ID listed, refers to Figure 1G.

Supplementary Figure 8. Immunostaining of tissue biopsies

A. Representative images of immunolabeling for CD4, Foxp3, nuclear stain (DAPI) and respective quantification of percentage of CD4⁺Foxp3⁺ and CD4⁺Foxp3^- T cells in paired biopsies. Scale bar:100 μm. B. Relative KRAS^G12D expression in Panc-1 cells treated with iExoKras^G12D stored for 4 years at –80°C. Data are presented as mean (A) and mean +/– SEM (B). Unpaired t test without Welch’s correction, **P<0.01.

Supplementary Figure 9. Contribution of CD4⁺ and CD8⁺ T cells in iExoKras^G12D therapy response in mice

A. Schematic of experimental timeline using orthotopic KPC-689 mice. B. Representative images and respective quantification of CD4⁺ T cells (CD4), and CD8⁺ T cells (CD8) by immunolabeling in control and iExoKras^G12D treated orthotopic KPC-689 mice, n=4 mice per group. Scale bar: 100 μm. C. Schematic of experimental timeline using KPC, KPC CD4^−/− and KPC CD8^−/− GEMs treated with iExoKras^G12D. D. Representative MRI images and quantification of percent change in tumor burden at indicated time points of KPC, KPC CD4^−/− and KPC CD8^−/− treated with iExoKras^G12D, n=5 to 8 mice per group. E. Kaplan-Meier survival curves showing disease specific survival of KPC-Control (n=15 mice), KPC-iExo (n=7 mice), KPC CD4^−/− control (n=9 mice), KPC CD4^−/− iExo (n=7 mice), KPC CD8^−/− control (n=10 mice), KPC CD8^−/− iExo (n=7 mice). F qPCR analysis of Kras^G12D, Fas and Il15 expression following treatment with iExoKras^G12D or scrambled (ScrExo, control) (2 doses, 24h apart). RNA extracted at 72h (n=3 wells per group). G ChIP assays demonstrating the relative binding of DNMT1, EZH2, H3K27me3, and H3K27ac on FAS promoter in KPC-689 cells in the indicated treatment groups (2 doses, 24h apart). n=4-5 biological replicates per group. Data are presented as mean +/– SD. Unpaired t test without Welch’s correction (B, F, G), one-way ANOVA with Tukey’s/Dunnett’s multiple comparisons test (D), log-rank test (E). *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; ns, not significant.

Supplementary Figure 10. TLS in iExoKras^G12D and ICB combination treatment

A. Representative H&E (A) and quantification (B) of TLS in the indicated groups. Scale bar: 100 μm. Data are presented as mean +/– SD. Kruskal-Wallis with Dunn’s multiple comparisons test. **P<0.01.