Proteolytic pan-RAS cleavage leads to tumor regression in patient-derived pancreatic cancer xenografts

Abstract

The lack of effective RAS inhibition represents a major unmet medical need in the treatment of pancreatic ductal adenocarcinoma (PDAC). Here, we investigate the anticancer activity of RRSP-DT_B, an engineered biologic that cleaves the Switch I of all RAS isoforms, in KRAS-mutant PDAC cell lines and patient-derived xenografts (PDXs). We first demonstrate that RRSP-DT_B effectively engages RAS and impacts downstream ERK signaling in multiple KRAS-mutant PDAC cell lines inhibiting cell proliferation at picomolar concentrations. We next tested RRSP-DT_B in immunodeficient mice bearing KRAS-mutant PDAC PDXs. Treatment with RRSP-DT_B led to ≥95% tumor regression after 29 days. Residual tumors exhibited disrupted tissue architecture, increased fibrosis and fewer proliferating cells compared to controls. Intratumoral levels of phospho-ERK were also significantly lower, indicating in vivo target engagement. Importantly, tumors that started to regrow without RRSP-DT_B shrank when treatment resumed, demonstrating resistance to RRSP-DT_B had not developed. Tracking persistence of the toxin activity following intraperitoneal injection showed that RRSP-DT_B is active in sera from immunocompetent mice for at least one hour, but absent after 16 hours, justifying use of daily dosing. Overall, we report that RRSP-DT_B strongly regresses hard-to-treat KRAS-mutant PDX models of pancreatic cancer, warranting further development of this pan-RAS biologic for the management of RAS-addicted tumors.

INTRODUCTION

Pancreatic cancer is the 11^th most commonly diagnosed cancer in the U.S. and the 3^rd leading cause of cancer-related deaths, with an overall 5-year survival rate of 9% ¹. Poor prognosis is mainly attributed to late-stage clinical detection, early metastatic dissemination and lack of effective treatment options ^{2, 3}. Pancreatic ductal adenocarcinoma (PDAC) is the most common histologic subtype of pancreatic cancer, accounting for ~90% of cases ^{4, 5}. PDAC genetic profiling has found that KRAS is mutated in ~95% of patients, with NRAS and HRAS mutations accounting for <1% each and wild-type KRAS (KRAS^WT) for the remaining 5% ^6-8. Mutations in RAS genes (H-, N- and KRAS) result in RAS proteins being locked in the active GTP-bound state with consequent permanent activation of downstream RAS/MEK/ERK signaling, which controls cancer cell proliferation and survival ^9-11. Following decades of failures, inhibition of KRAS specific subtypes has had recent success ^12-16. Notably, sotorasib (LUMAKRAS™) was recently approved by the FDA as the first therapy for patients with KRAS^G12C -mutant non–small cell lung cancer ^{17, 18}.

However, the usefulness of these KRAS^G12C inhibitors for PDAC is limited to ~2% of patients ¹⁹. Therefore, there is an urgent need for effective therapeutic strategies that target KRAS directly in PDAC.

We recently demonstrated that the pan-RAS biologic RRSP-DT_B inhibits tumor growth in xenograft models of triple-negative breast cancer (TNBC) and colorectal cancer (CRC) in mice via direct RAS cleavage ²⁰. RRSP-DT_B is an engineered chimeric toxin comprised of the RAS/RAP1 specific endopeptidase (RRSP) from Gram-negative Vibrio vulnificus and the protein delivery machinery of diphtheria toxin (DT_B). DT_B delivers RRSP intracellularly via Heparin-binding EGF-like growth factor (HB-EGF)-mediated endocytosis in receptor-bearing cells only ²⁰. Human HB-EGF acts as the unique receptor of diphtheria toxin (DT) and is widely expressed in epithelial tumors, including PDAC ²¹, where it has been shown to cooperate with KRAS to promote KRAS-driven tumorigenesis ²². Once released into the cytoplasm, RRSP cleaves RAS and RAP1 with high specificity within the Switch I region, which enables RAS binding to effectors and downstream signal transduction ^{23, 24}. All three RAS isoforms (H-, N- and KRAS) are cleaved by RRSP whether they are bound to GDP or GTP, so that the entire RAS cellular pool is inactivated ²⁵. Importantly, RRSP processes the main oncogenic RAS proteins with activating point mutations at residues G12, G13 and Q61 ²³. Cleavage of all RAS isoforms, mutant and wild-type, is an inherent advantage of RRSP and represents a potent strategy to reduce or regress growth in a broader spectrum of tumor types. The processing of RAS by RRSP-DT_B leads to a variety of cellular outcomes including cell cycle arrest via upregulation of p27, apoptosis, or senescence. The differential cell fates were not correlated with the KRAS mutation, but rather, the mechanisms for inhibition of cell growth upon loss of RAS depend on signaling networks downstream of RAS ²⁶.

Here, we use human KRAS-mutant PDAC cell lines and clinically relevant PDAC patient-derived xenografts (PDXs) to demonstrate preclinical effectiveness of the pan-RAS inhibitor RRSP-DT_B against hard-to-treat PDAC. We also provide data on the stability of the toxin in the bloodstream following intraperitoneal (i.p.) injection, useful to support advanced RRSP-DT_B therapeutic development.

MATERIALS AND METHODS

Protein preparations, chemicals and cell lines

Purification of RRSP-DT_B and catalytically inactive RRSP*-DT_B proteins and endotoxin removal were performed as previously described ²⁰. All chemicals were from Sigma-Aldrich unless otherwise specified. Validated cell lines were kindly provided from the Reference Reagent Resource of the National Cancer Institute RAS Initiative at Frederick National Laboratory for Cancer Research (FNLCR). Cell lines at this resource are obtained by the Ras Initiative from collaborators or commerical sources and confirmed free of Mycoplasma using VenorGeM Mycoplasma Classic Endpoint PCR assay and are also subjected to short tandem repeat analysis using the AmpFLSTR Identifiler PCR Amplification Kit to authenticate the cell lines, comparing the results to information located at https://web.expasy.org/cellosaurus/. Cells were cultured at 37°C and 5% CO₂ atmosphere. YAPC and SUIT-2 were grown in RPMI-1640 (ATCC formulation) containing 10% fetal bovine serum (FBS; Gemini) and 1% penicillin/streptomycin (P/S; Invitrogen). KP-1N were grown in RPMI-1640 containing 5% FBS and 1% P/S. KP-4 were grown in DMEM-F12 with Glutamax (Gibco) containing 5% FBS and 1% P/S. PANC-1 were grown in DMEM (ATCC formulation) with 10% FBS and 1% P/S.

Cytotoxicity, viability, clonogenic and apoptosis assays

Cytotoxicity was assessed by staining cells with crystal violet. Briefly, 50,000 cells/well were cultured in 48-well plates and treated with RRSP-DT_B and RRSP*-DT_B for 72 hours. Cells were washed and crystal violet fixing/staining solution was added for 20 min at room temperature as previously described ²⁰. Images of air-dried plates were acquired using a conventional desktop scanner.

Cell viability was measured using CellTiter-Glo (Promega). 10,000 cells/well were grown in 96-well clear bottom white plates and treated with RRSP-DT_B and RRSP*-DT_B for 72 hours. CellTiter-Glo was then added to each well following the manufacturer’s instructions and luminescence was recorded using a Tecan Safire2 plate reader. IC₅₀ were calculated using the log(inhibitor) vs. response - variable slope (four parameters) function in Graphpad Prism v8.

Cell proliferation was assessed by colony-formation assay. RRSP-DT_B and RRSP*-DT_B-treated cells for 72 hours were trypsinized, counted on a hemocytometer and replated in 48-well plates at 250 cells per well. Culture medium was replaced biweekly. On day 14, a crystal violet fixing/staining solution was added to the plates. Quantitative changes in cytotoxicity or clonogenicity were determined by solubilizing crystal violet-stained plates with 10% acetic acid and measuring the absorbance at 590 nm using a plate reader (Tecan Safire2).

Apoptosis was assessed by Caspase-Glo 3/7 assay (Promega). 10,000 cells/well were grown in 96-well clear bottom white plates and treated with 0.1 and 10 nM of RRSP-DT_B and RRSP*-DT_B as well as 1 µM of Staurosporine (positive control) for 24 and 48 hours. Caspase-Glo 3/7 was then added to each well following the manufacturer’s instructions and luminescence was recorded using a Tecan Safire2 plate reader.

SDS-PAGE and western blotting

Protein extracts were prepared from cells by adding M-PER mammalian protein extraction reagent (Thermo Fisher Scientific) with protease and phosphatase inhibitors (Sigma-Aldrich). Protein content was measured using the Bio-Rad protein assay dye reagent concentrate (#5000006). Equal amounts of proteins were separated by SDS-PAGE followed by western blot analysis as previously described ²⁴. Membranes were blotted using the following antibodies: anti-panRAS (Thermo Fisher Scientific Cat# MA1–012, RRID:AB_2536664), which recognizes RAS Switch I and thus detects only uncleaved RAS. Anti-mAb 4E8, a pan-RAS monoclonal antibody that recognizes both cleaved and uncleaved bands of all three RAS isoforms was prepared from a hybridoma cell line obtained from the National Cancer Institute RAS Initiative. The antibody preparation used in this study was previously validated ²⁰. Anti–Phospho-p44/42 MAPK (pERK1/2; Cell Signaling Technology Cat# 4377, RRID:AB_331775), anti-p44/42 MAPK (ERK1/2; Cell Signaling Technology Cat# 4696, RRID:AB_390780), and anti-HB-EGF (Abcam, #ab185555). Anti-vinculin (Cell Signaling Technology Cat# 13901, RRID:AB_2728768) was used for normalization. Secondary antibodies used were fluorescent-labeled IRDye 680RD goat anti-mouse (LI-COR Biosciences Cat# 926–68070, RRID:AB_10956588) and IRDye 800CW goat anti-rabbit (LI-COR Biosciences Cat# 926–32211, RRID:AB_621843). Blot images were acquired using the Odyssey Infrared Imaging System (LI-COR Biosciences) and quantified by densitometry using NIH ImageJ software (ImageJ, RRID:SCR_003070). Percentage of uncleaved RAS was calculated as previously described ²⁴.

Patient-derived xenograft (PDX) studies

Human PDAC biospecimens were obtained from the Pathology Core Facility (PCF) Biobank at Northwestern University. Collection of human tissue specimens and the PDX mouse protocol were approved by the Institutional Review Board and Institutional Animal Care and Use Committee (IACUC) at Northwestern University. All PDAC tumor models used have been fully tested and validated across the entire spectrum of creating, propagating, and characterizing the PDX models. The established PDAC tumor models are defined as the surgical collected fresh tumor tissue are subcutaneously implanted into NSG mice. This process is repeated two more times to ensure that the tumors can be propagated through at least three passages. Cryopreserved fragments are thawed and re-implanted to make sure that frozen tissue will regrow. At each passage, tumor tissue is compared to the original patient tumor on the basis of histopathology, immunohistochemistry for expression of clinically relevant biomarkers. Only PDX tumors that retain the histological characteristics of the patient tumor and to exclude development of lymphoma are considered confirmed. All tissues are routinely tested for Mycoplasma using the Universal Mycoplasma Detection Kit (American Type Culture Collection, Cat #30–1012K) for our studies.Molecular profiling for tumor genomic and transcriptomic characteristics and response to a standard-of-care combination therapy for PDAC PDX tumor are also used to establish the baseline characteristics of the model. At the time of this study, multiple KRAS^G12V PDXs were available from the repository and we selected two of them from patients that did not receive any adjuvant or neo-adjuvant therapy highlighting the different sex (female vs male) as a valuable variable. Although our repository included one KRAS^G12D PDX, it was from a patient that received multiple neo-adjuvant therapies (gemcitabine, oxaliplatin), tumor pathology/staging report was not available at the time of the study and therefore was not considered representative for this study. PDXs were established by the Center for Developmental Therapeutics (CDT) at Northwestern University ²⁷. Briefly, cryopreserved human PDAC biospecimens (passage 0, P₀) were thawed, cut into small fragments, and two tissue fragments were xenografted into the dorsal region of two female NSG mice (P₁). When tumor masses reached ~1500 mm³ in size, tumors were excised, divided into equally sized specimens (~2 × 2 × 2 mm) and implanted subcutaneously into the dorsal region of the appropriate number of female NSG mice (P₂). Once tumors reached 150–200 mm³, mice underwent balanced randomization based on tumor size. Briefly, mice were grouped three per cage (two cages used for each treatment) so that differences in the median tumor size were minimized among groups prior to treatment administration. Next, mice were dosed i.p. with endotoxin-free 0.1 mg/kg of RRSP-DT_B or RRSP*-DT_B on a 5 days ON/2 days OFF schedule for four weeks. Control mice received endotoxin-free phosphate buffered saline (MilliporeSigma). After four weeks, all mice (except three mice from the RRSP-DT_B-treatment group) were humanely euthanized, tumors excised, and fixed in 10% formalin overnight. The remaining three mice from the RRSP-DT_B group were housed without treatment for three weeks and subsequently dosed with RRSP-DT_B at the indicated treatment regimens. Tumor growth was monitored by digital caliper measurements of length (l) and width (w) of the tumors. Tumor volume was calculated using the following formula: volume (mm³) = (l x w²)/2. Mouse body weight was also measured regularly. Percentage change in tumor volume between day 1 and day 29 was calculated for each individual tumor as described in Patel et al. ²⁸

In vivo stability study

Five female C57BL/6 mice per group were injected i.p. with 0.1 mg/kg or 0.5 mg/kg of RRSP-DT_B. Saline was used as vehicle control. After 1 and 16 hours, blood was collected from anesthetized mice via orbital bleeding, transferred to serum separator collection tubes (BD Biosciences, #365967), allowed to clot for at least 30 minutes and then centrifuged for 10 minutes at 10,000xg. The serum supernatant was aliquoted and stored at −80°C until use. To determine whether RRSP-DT_B was active in the blood after a single injection, serum collected from mice was diluted 1:100 dilution with saline, added directly to YAPC or KP-4 cultured pancreatic cells, and total levels of intracellular RAS cleaved after 24 hours was assessed by western blot as described above.

Histology, immunohistochemistry and image analysis

Paraffin-embedding, sectioning, hematoxylin and eosin (H&E) and Masson’s trichrome stainings as well as immunohistochemical staining of PDX specimens were performed by the Robert H. Lurie Comprehensive Cancer Center Pathology Core Facility. Tumor sections were stained with anti-Cytokeratin 19 ((CK-19), #ab76539; Abcam), anti-Ki-67 (#GA626; Dako), anti-cleaved caspase 3 ((CC3, #9661; Cell Signaling Technology), anti-pan-RAS ((RAS), #PA5–85947; Thermo Fisher Scientific) and anti-Phospho-p44/42 MAPK ((ERK1/2) (Thr202/Tyr204, (D13.14.4E) XP, #4370; Cell Signaling Technology) antibodies as described previously ²⁰. Primary antibodies were detected using the appropriate secondary antibodies and 3,3′-diaminobenzidine (DAB) revelation (Dako).

All slides were scanned using a Nanozoomer HT slide scanner (Hamamatsu). Quantification of positively immunostained cells was performed using customized Application Protocol Packages (APPs) within the VIS image analysis software (Visiopharm). Positive cells were counted in the whole tumor sections and expressed as the percentage ratio over the area of the whole section. Because the pattern of cytoplasmic RAS staining was diffuse, we quantified RAS signal intensity by color deconvolution using ImageJ (Fiji version) as previously described ²⁹.

Statistical analysis

Graphpad Prism v.8 software was used for statistical analysis. Bar plots represent the mean of at least three independent experiments ± the standard deviation (SD). Statistical significance was assessed using one-way analysis of variance (ANOVA) assuming normal distribution. Dunnett’s multiple comparison post-test was employed to compare the mean of the control group to the mean of treatment groups. Tukey’s multiple comparison test was used to compare the mean of each group with the mean of every other group. Points in the fitted dose-response curve represent mean ± standard error of the mean (SEM). Statistical analysis on fold-change data was performed after log transformation to obtain a more normalized distribution. For in vivo PDXs, data are reported as mean ± SEM and one-way ANOVA was performed to assess differences among the treatment arms. Values of p<0.05 were considered statistically significant.

Data Availability

The data generated in this study are available within the article and its supplementary data files.

RESULTS

RRSP-DT_B cleaves RAS in KRAS-mutant PDAC cell lines

For this study, we selected five different KRAS mutant PDAC cell lines, including YAPC (G12V), SUIT-2 (G12D), KP-1N (G12D), KP-4 (G12D) and PANC-1 (G12D). All five cell lines used expressed the DT receptor HB-EGF. However, an originally selected cell line HUP-T4 cells expressed low levels of HB-EGF, and subsequently was excluded from the study (Supplementary Figure S1A).

In order to assess intracellular delivery of RRSP via DT_B and successful RAS target engagement, PDAC cells were treated with increasing concentrations of RRSP-DT_B up to 10 nM and with 10 nM of catalytically inactive RRSP*-DT_B for 24 hours. Western blotting of cell lysates showed that RRSP-DT_B cleaved total RAS in all PDAC cell lines with similar picomolar potency (Figure 1A-E). The half-maximal inhibition concentration (IC₅₀) values extrapolated from dose-response curves of uncleaved RAS normalized to vinculin showed that 50% of RAS cleavage was achieved with concentrations ranging between 10 and 280 pM (Figure 1F). The appearance of cleaved RAS also tracked with a significant reduction in levels of phosphorylated ERK (pERK) in all five cell lines (Figure 1A-F). Altogether, these data demonstrate successful DT_B-mediated delivery of RRSP in PDAC cells and effective target engagement as shown by RAS cleavage and reduced pERK.

Figure 1: Assessment of RAS cleavage by RRSP-DT_B in KRAS-mutant PDAC cell lines.

(A-E) Representative western blots and quantification of uncleaved/cleaved RAS and pERK levels in (A) YAPC, (B) SUIT-2, (C) KP-1N, (D) KP-4 and (E) PANC-1 PDAC cell lines following treatment with increasing doses of RRSP-DT_B as indicated or 10 nM of catalytically inactive mutant RRSP*-DT_B for 24 hours. Bar plots represent mean ± SD of three independent experiments (*p < 0.05, **p < 0.01, ****p < 0.0001 versus control 0 nM; one-way ANOVA followed by Dunnett’s multiple comparison test; n= 3). Superimposed solid red dose-response curves were used to extrapolate the concentration of RRSP-DT_B required to cleave 50% of RAS in the corresponding cell line, as reported in (F).

RRSP-DT_B impacts viability and proliferation of KRAS-mutant PDAC cell lines

To examine the cytotoxic and/or growth inhibitory effects of RRSP-DT_B, the treatment of PDAC cell lines was extended to 72 hours. A crystal violet cytotoxicity assay showed reduced confluency for YAPC and SUIT-2 cells indicating cytotoxicity, and moderate cell loss in KP-1N, KP-4 and PANC-1 cells compared to the control treatments indicating growth inhibition (Figure 2A and Supplementary Figure S1B). Quantitative assessment of cell viability using CellTiter-Glo, which detects metabolic activity, confirmed that YAPC and SUIT-2 cells lost viability with IC₅₀ in the picomolar range, while KP-1N, KP-4, and PANC-1 remained viable although with reduced proliferation (Figure 2B).

Figure 2: Effect of RRSP-DT_B on viability and proliferation of KRAS-mutant PDAC cell lines.

(A) Representative image of crystal violet-stained YAPC, SUIT-2, KP-1N, KP-4 and PANC-1 cells following 72 hours of treatment with varying doses of RRSP-DT_B as indicated and 10 nM of RRSP*-DT_B for 72 hours. (B) Dose–response curves showing the effect of RRSP-DT_B and RRSP*-DT_B on viability of multiple PDAC cell lines following 72 hours of treatment and summary of extrapolated IC₅₀ values. Results are expressed as mean ± SEM (n= 3). (C) Dose–response curves of RRSP-DT_B and RRSP*-DT_B on 3D cultures of KP-4 cells and representative pictures of corresponding spheroids. (D) Representative image of crystal violet-stained colonies from YAPC, SUIT-2, KP-1N, KP-4 and PANC-1 cells pretreated with RRSP-DT_B and RRSP*-DT_B for 72 h and replated at low density to form colonies over 14 days. (E) Quantification of colonies shown in (D) from three independent experiments. Results are expressed as means ± SD (**p < 0.01, ****p < 0.0001 versus corresponding control 0 nM; one-way ANOVA followed by Dunnett’s multiple comparison test, n= 3).

KP-4 cells were found to generate structurally well-defined three-dimensional (3D) spheroids. When treated with RRSP-DT_B, these spheroids not only showed a strong dose- and time-dependent reduction of size, but also of viability as measured by the 3D CellTiter-Glo assay (Figure 3C). These data show that the KRAS-mutant KP-4 cell line became more sensitive to RAS inhibition in 3D spheroid culture conditions compared to monolayers, consistent with previous findings ³⁰.

Figure 3: Evaluation of the in vivo activity of RRSP-DT_B in KRAS-mutant PDAC PDXs.

(A, B) Average growth curves of KRAS^G12V PDAC PDX tumors from a female donor (A, PDX1) and a male donor (B, PDX2) in NSG mice treated with vehicle (saline), 0.1 mg/kg of RRSP-DT_B and 0.1 mg/kg of RRSP*-DT_B. Mice were injected intraperitoneally every day (weekends excluded) for four weeks followed by a drug-washout phase of 24 days (no treatment). Drug treatment was resumed on day 24 in three mice from PDX1 that received 0.1 mg/kg of RRSP-DT_B q.o.d during the first week and q.d. during the second week following washout (A). In PDX2, mice received 0.1 mg/kg of RRSP-DT_B q.d. for four additional weeks after drug washout (B). Dotted lines indicate the average tumor volume on the first day of treatment. In PDX1, grey data point in Figure 3B and grey bar in Figure 3C are from a mouse that was euthanized earlier because of excessive tumor burden compare to its weight. (C, D) Representative images of PDX1 (C) and PDX2 (D) tumors at day 29 and corresponding column scatter plots showing individual tumor volumes. Data are means ± SEM (n= 6 mice per group; **p < 0.01, ****p < 0.0001; one-way ANOVA followed by Tukey’s multiple comparison test, n = 6). (E, F) Waterfall plots depicting tumor regression as percentage change in tumor size from baseline for individual mice.

The sensitivity of all cells to RRSP-DT_B resulting in either cytotoxicity or irreversible growth inhibition was confirmed using a clonogenic assay. Following RRSP-DT_B treatment for 72 hours, cells were harvested, counted, and replated at low density. At day 14, all cell lines failed to form colonies equivalent to the control treatment groups demonstrating an inability of residual cells to proliferate. Overall, these data reveal that RRSP-DT_B strongly affects viability and proliferation of KRAS-mutant PDAC cell lines. Importantly, cell lines previously classified as KRAS-dependent were confirmed to be highly susceptible, and those classified as KRAS-independent (KP-1N, KP-4 and PANC-1) ³¹ were also sensitive to RRSP-DT_B treatment.

RRSP-DT_B leads to in vivo regression of KRAS-mutant PDAC xenografts

Patient-derived xenografts (PDXs) are a powerful tool in translational cancer research because they retain biological characteristics of the parental tumor specimens and can better predict a patient response to an investigational drug than classic cell line-based in vivo models of cancer ³². We established two PDAC PDX models, one from a female donor (PDX1) and one from a male donor (PDX2), both harboring a KRAS G12V mutation, which accounts for 30% of KRAS mutations in PDAC ²⁷. NSG mice were engrafted with patient-derived PDAC tumors (eighteen mice per PDX group) and then six mice per group were treated i.p. with 0.1 mg/kg of RRSP-DT_B or RRSP*-DT_B once per day (q.d.) for 4 weeks (weekends excluded) or were mock-injected with saline. RAS inhibition by RRSP-DT_B significantly reduced the growth of subcutaneously implanted KRAS-mutant PDX1 and PDX2 tumors, starting as soon as 8 days after treatment initiation (Figure 3A and 3B). After four weeks of treatment, on day 29, maximum growth inhibition was observed (Figure 3C and 3D). In both PDX1 and PDX2 studies, tumor regression, not just inhibition of tumor growth, was observed in all twelve mice treated with RRSP-DT_B indicating successful treatment of pre-established tumors (Figure 3E and 3F). The reduction of tumor size was specifically due to the proteolytic action of RRSP, since the catalytically inactive RRSP*-DT_B did not affect tumor growth.

In order to determine whether tumor growth occurred in residual tumors after stopping the treatment, three mice from the RRSP-DT_B group were left untreated from day 29 for about three weeks and tumor growth was regularly monitored. In both PDX1 and PDX2, we did not observe an increase in tumor size during the first 10 days (day 29–39) following RRSP-DT_B withdrawal. Between day 39 and day 53, a slight increase in tumor size occurred in mice from PDX1 (Figure 3A) and PDX2 (Figure 3B), where the average tumor size reached the initial baseline. On day 53, treatment with RRSP-DT_B was reinitiated. The treatment of PDX1 mice was originally de-escalated to every other day (q.o.d). As there was no reduction in tumor size after one week, RRSP-DT_B treatment frequency was increased to five days per week during the second week. Increase in the treatment frequency ultimately led to a reduction in tumor size (Figure 3A).

Since tumors from PDX2 had a higher growth rate compared to those from PDX1, mice were retreated with RRSP-DT_B once per day (q.d.) for four weeks. At the end of this additional four-week treatment cycle, tumors reached a size comparable to that observed around day 39 (Figure 3B). Of importance, RRSP-DT_B treatment was well tolerated in mice, as demonstrated by the lack of weight loss (Supplementary Figure S2) across both treatment cycles.

Altogether, these data demonstrate that RRSP-DT_B treatment resulted in rapid in vivo regression of PDAC PDXs. Most importantly, PDAC tumors retained sensitivity to RRSP-DT_B over time, did not develop resistance and remained sensitive for a second round of treatment.

RRSP-DT_B treatment results in increased fibrosis, reduced proliferation and downregulation of pERK in in vivo PDAC tumors.

Histological and immunohistochemical analysis showed that the large PDX tumors resected from vehicle- and RRSP*-DT_B-treated mice were characterized by small glandular components and extensive necrosis in the inner tumor core typical of clinical PDAC. By contrast, tumors from mice treated with RRSP-DT_B showed absence of necrosis, reduction in the cell number, larger residual tumor cells and appearance of vacuolar structures indicative of tissue degeneration (Supplementary Figure S3). In addition, RRSP-DT_B-treated tumors were highly fibrotic (Figure 4A and 4B).

Figure 4: Histological and immunohistochemical analysis of PDAC PDX tumors following treatment with RRSP-DT_B.

(A, B) Representative images of saline, RRSP*-DT_B and RRSP-DT_B-treated PDAC PDX1 (A) and PDX2 (B) tumors stained with H&E and Masson’s trichrome stain as well as CK-19, Ki-67 and pERK antibodies (scale bar = 250 µm). (C-E) Bar plots represent cell positivity to CK-19, Ki-67 and pERK antibodies expressed as the percentage of immunoreactive cells in the entire tumor sections in PDX1 and (F-H) PDX2. Data are means ± SEM (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus saline; #p < 0.05, ##p < 0.01, ###p < 0.001 versus RRSP*-DT_B; one-way ANOVA followed by Tukey’s multiple comparison test, n = 3).

Cytokeratin 19 (CK-19) is a characteristic ductal epithelial marker and poor prognostic factor in PDAC ^{33, 34}. The percentage of cells in the tumors that stained positive for CK-19 was significantly lower in all tumors treated with RRSP-DT_B compared to saline and RRSP*-DT_B-treated mice (Figure 4C and 4F). This result indicated that fewer PDAC cells were present in the residual tumors. In addition, tumors from mice in the RRSP-DT_B treatment group showed a marked reduction in proliferating Ki-67 positive cells, both after the first 4 weeks of treatment and also after the second treatment regimen that started at day 53 (Figure 4D and 4G).

In addition, PDAC tumors from PDX1 and PDX2 excised at day 29 showed significantly lower detection of pERK compared to saline and RRSP*-DT_B-treated tumors (Figure 4E and 4H). In PDX1, we observed an increase in pERK levels on day 67 in tumors from mice that received RRSP-DT_B every other day (q.o.d) during the first week and once per day (q.d.) during the second week following RRSP-DT_B withdrawal. This tracked with a higher percentage of residual cancer cells positive to CK-19 and Ki-67 compared to day 29. By contrast, in mice from PDX2 that received the more aggressive second round of treatment, low levels of pERK were also observed in tumors resected on day 81. Analysis of total RAS levels showed significant reduction of RAS expression in sections from tumors treated with RRSP-DT_B at day 67 in PDX1 and, although not significant, a similar trend was observed in PDX2 at day 29 (Supplementary Figure S4). RAS levels in histological sections from PDX2 tumors at day 81 are not shown because of staining-related technical issues.

Next, to investigate whether apoptosis was involved in RRSP-DT_B-mediated anti-tumor activity, cleaved caspase 3 (CC3) expression was assessed in both PDX1 and PDX2. No changes in CC3 levels were detected in tumors treated with RRSP-DT_B compared to controls. Similarly, activation of caspase 3/7 was not observed in four out of five PDAC cell lines tested. Only YAPC cells showed an increase of caspase 3/7 activity upon treatment with RRSP-DT_B (Supplementary Figure S5).

Taken together, these results demonstrate that treatment of PDAC PDXs with RRSP-DT_B leads to significant reduction in the percentage of cancer cells, increased fibrosis and decreased proliferation independently of apoptosis. Importantly, RRSP-DT_B strongly reduces pERK levels, which is indicative of effective downstream RAS signaling downregulation.

Evaluation of the in vivo stability of RRSP-DT_B in immunocompetent mice.

An important aspect underlying the translatability of RRSP-DT_B as an anticancer agent is to determine its in vivo stability. In order to do this, immunocompetent C57/BL6 mice were given a single i.p. injection of 0.1 or 0.5 mg/kg of RRSP-DT_B. Serum from mice collected one and 16 hours after injection was diluted 1:100 and then was added to YAPC and KP-4 cells to test for recovery of cellular active RRSP-DT_B from the bloodstream.

We found that in sera from mice that received 0.5 mg/kg of RRSP-DT_B for one hour, RRSP-DT_B was active. Indeed, a remarkable reduction in RAS levels was observed in both YAPC and KP-4 cells treated with mouse sera. In KP-4 cells, reduced RAS levels were also detected at 0.1 mg/kg of RRSP-DT_B. We did not observe a significant decrease of RAS levels in cells treated with sera from mice that received RRSP-DT_B for 16 hours (Figure 5A and Supplementary Figure S6), suggesting that RRSP-DT_B that had been exposed to mouse serum for a longer time became less active and was not able to significantly reduce RAS levels as efficiently as at earlier times.

Figure 5: Stability of RRSP-DT_B in sera from immunocompetent mice.

(A) Densitometric quantification of total RAS levels in KRAS-mutant YAPC and KP-4 cells treated with serum from mice that were previously injected with 0.1 mg/kg or 0.5 mg/kg of RRSP-DT_B for 1 hour and 16 hours. Each bar represents the mean ± SEM of RAS levels from 5 different mice per group (*p < 0.05, ***p < 0.001 versus saline; one-way ANOVA followed by Dunnett’s multiple comparison test, n= 5).

DISCUSSION

Aberrant KRAS signaling is considered the core hallmark of pancreatic cancer onset and progression ³⁵. Although significant advances have been made over recent years in the treatment of pancreatic cancer ³⁶, there are currently no KRAS inhibitors approved for PDAC, which still remains an urgent unmet medical need in oncology.

We recently reported the therapeutic potential of biologic RRSP-DT_B as a potent pan-RAS inhibitor that halts tumor growth of multiple TNBC and CRC xenografts harboring either wild-type or mutant RAS ³⁷. The NCI-60 human tumor cell line screening revealed that RRSP-DT_B, in addition to its potency, has a broad spectrum of antitumor activity against several cancer types ²⁰. Because the NCI-60 screen does not include pancreatic cancer, we tested RRSP-DT_B on a panel of five KRAS-mutant PDAC cell lines and demonstrated that picomolar concentrations of RRSP-DT_B leads to growth inhibition and loss of cell proliferation, although only cells treated with the highest dose of 10 nM showed loss of metabolic activity. This was not completely unexpected since KP-1N, KP-4 and PANC-1 cells were previously classified as KRAS-independent by Singh and co-authors ³¹. Interestingly, when KRAS-independent KP-4 cells were cultured as 3D spheroids, a stronger decrease in viability and spheroids’ size was observed compared to monolayers. We believe the greater sensitivity of KP-4 spheroids to RRSP-DT_B might be due to the 3D suspension culture environment, thus corroborating previous findings showing that different culture systems have an impact on KRAS dependency ³⁰. These data confirm that PDAC cell lines previously listed as KRAS-dependent were extremely susceptible to RRSP-DT_B. Most importantly, long-lasting growth inhibitory effects were seen in mutant KRAS-independent PDAC cell lines as well. Although the mechanisms underlying RRSP-DT_B-induced growth inhibition in PDAC cell lines are still unknown, we have recently demonstrated that RRSP-DT_B can lead to irreversible G1 cell cycle growth arrest via p27, induce a senescence-like phenotype and trigger apoptosis in a subset of colorectal cancer cell lines ³⁸. Our data show that apoptosis is not overall activated in PDAC tumors or cell lines following RRSP-DT_B treatment. Although this was surprising, it was not completely unexpected since, in a bacterial pathogenesis context, toxins such as RRSP have evolved to hijack cellular pathways in order to avoid host innate responses and promote systemic spread without killing of the host. Overall, our data suggest that the main mechanisms underlying RRSP’s antitumor activity involve cell cycle arrest and/or senescence rather than apoptosis ²⁶. In support of our findings, an independent study has recently shown that RRSP from Photorhabdus asymbiotica, which is 73% similar in amino acid sequence to RRSP from Vibrio vulnificus, also cleaves RAS ²⁴ and induces G1 cell cycle arrest via direct binding to CDK1 ³⁹.

Our data suggest that, unlike currently available KRAS inhibitors, RRSP-DT_B has the potential to target undruggable KRAS in a wide spectrum of KRAS-mutant pancreatic cancers whether they rely or not on KRAS for survival. One of the inherent advantages of RRSP is that it cleaves all RAS isoforms, mutant and wild-type, thus representing a potent strategy to inhibit tumor growth. While one line of research has been focusing on targeting mutant KRAS only with molecules that enter all cells to ensure safety, our current efforts are directed to transporting RRSP preferentially into cancer cells while sparing normal cells, an approach previously found to be successful for other toxin-based targeting strategies ^{40, 41}, including the therapeutic Lumoxiti^®, which is an FDA-approved treatment for hairy cell leukemia ⁴².

RRSP-DT_B has been shown to halt tumor growth in mouse xenografts derived from TNBC and CRC cell lines ³⁷. Here, we report that administration of RRSP-DT_B to NSG mice bearing KRAS-mutant PDAC PDXs not only slowed tumor growth, but also led to tumor regression as early as one week after treatment initiation. Following a drug washout period, tumors from both PDX1 and PDX2 started to slowly regrow. Most importantly, upon re-administration of RRSP-DT_B, these tumors maintain sensitivity to the drug without becoming resistant, which is something rarely seen in monotherapy regimens and might be advantageous for future development of RRSP as an anticancer therapeutic. Lack of resistance could also be explained by the fact that at day 29 residual tumors are mainly composed by a dense network of collagen fibers with few interspersed CK-19-positive PDAC cancer cells that can be readily targeted by RRSP-DT_B. PDAC tumors have extensive fibrosis, which is believed to contribute to their resistance to most therapies. We showed that RRSP-DT_B increased intracellular fibrosis in vivo, which could be considered disadvantageous. However, contrary to expectations, experimental ^{43, 44} and clinical ⁴⁵ studies have demonstrated that depletion of the stroma component of PDAC results in more aggressive tumors in mice and worse outcomes in patients. Therefore, rather that depleting the stroma, strategies that specifically target or modify the extracellular matrix (ECM) are preferred and currently being investigated in PDAC. Although we do not have a definitive answer to the question of whether RRSP-mediated increase in fibrosis is advantageous or disadvantageous following RRSP-DT_B treatment, based on the available literature and our findings, we can speculate that it will be beneficial. In addition, we believe that because DT_B has been shown to be extremely efficient at delivering RRSP in cells expressing the HB-EGF receptor, the non-cellular ECM component present in the tumor does not represent an obstacle for RRSP-DT_B to reach the residual cancer cells interspersed within the ECM, effectively inhibiting their growth.

In PDX1, because mice were subjected to a shorter RRSP-DT_B treatment regimen following the drug-washout phase, we observed a larger number of positive CK-19 cells and increase pERK levels despite no increased proliferation. These findings have helped us improve RRSP-DT_B treatment schedule. Indeed, a more aggressive dosing regimen as in PDX2 resulted in fewer CK-19-positive cells and reduced pERK expression comparable to those seen before RRSP-DT_B washout phase. Moreover, the in vivo stability study showed that RRSP-DT_B injected i.p. is delivered to the bloodstream in immunocompetent mice although with minimal activity persisting until 16 hours, supporting the more aggressive daily dosing schedule. Notably, despite the extended and more frequent in vivo administration of RRSP-DT_B, no systemic toxicity was observed in mice.

Altogether, our study provides compelling evidence that engineered RAS endopeptidase RRSP-DT_B is highly efficacious against PDAC. Another as yet untested potential advantage of RRSP-DT_B is that in addition to RAS, RRSP cleaves the metastasis-associated GTPase RAP1. Future studies in metastatic cancer models could be done to probe RRSP efficacy against metastatization and local invasion as an added advantage of RRSP in the treatment of PDAC,

In recent years, protein degradation has become a promising therapeutic modality to target RAS. Besides RRSP, the bacterial protease toxin subtilisin was recently re-engineered to degrade RAS in vitro and degraded RAS also when ectopically expressed in mammalian cells ⁴⁶. In addition, several studies showed that degradation of endogenous KRAS can be achieved following ectopic expression in cancer cells of chimeric proteins that tether RAS-targeting peptides ^47-50 and monobodies ⁵¹ to a E3 ligase that ultimately drives KRAS ubiquitination and proteasomal degradation. RAS degradation, unlike inhibition of the protein activity, is a strategy that may lead to prolonged inactivation of downstream RAS signaling bypassing the problem of intrinsic resistance of RAS proteins that have been covalently inhibited with small molecules. We likewise showed in 2015 that RRSP can cleave RAS when ectopically expressed resulting in loss of activity and reduced RAS levels ²³. In contrast to all other studies on engineered RAS degraders that depend on stable transfection and ectopic expression, RRSP-DT_B is the only RAS degrader shown to cleave RAS when added exogenously to cells and to reduce tumors following injection in mice. Further, we found that the active toxin moves from the peritoneum to the bloodstream. Thus, RRSP-DT_B is a first-in-class RAS degrader now shown herein to have potential to treat PDAC resulting in tumor regression. In contrast to small molecule inhibitors that target only one form of RAS ⁵², treatment with RRSP-DT_B showed no resistance in mice that received RRSP-DT_B after a drug washout phase. Because of the unique way RRSP readily cleaves and degrades RAS proteins and the fact that, to our knowledge, RRSP-DT_B is the most advanced and documented RAS biodegrader as for in vitro potency and in vivo activity, we believe that it is the leading player in the growing targeted RAS degradation space. Thus, further investigation of RRSP-DT_B as a therapeutic tool for degrading RAS is warranted and optimization of the delivery modality is planned to achieve selective cancer cell receptor targeting.