Tumor-Colonizing E. coli Expressing Both Collagenase and Hyaluronidase Enhances Therapeutic Efficacy of Gemcitabine in Pancreatic Cancer Models

Abstract

Desmoplasia is a hallmark feature of pancreatic ductal adenocarcinoma (PDAC) that contributes significantly to treatment resistance. Approaches to enhance drug delivery into fibrotic PDAC tumors continue to be an important unmet need. In this study, we have engineered a tumor-colonizing E. coli-based agent that expresses both collagenase and hyaluronidase as a strategy to reduce desmoplasia and enhance the intratumoral perfusion of anticancer agents. Overall, we observed that the tandem expression of both these enzymes by tumor-colonizing E. coli resulted in the reduced presence of intratumoral collagen and hyaluronan, which likely contributed to the enhanced chemotherapeutic efficacy observed when used in combination. These results highlight the importance of combination treatments involving the depletion of desmoplastic components in PDAC before or during treatment.

Keywords: bacterial vector; collagen; collagenase; desmoplasia; fibrosis; hyaluronan; hyaluronidase; pancreatic ductal adenocarcinoma; resistance.

Figure 1

Tandem construct design and expression analysis in BL21 E. coli transformants. (A) The synthesized sequence encoding for both bacterial collagenase (ColH) and hyaluronidase (HylB) was engineered with independent 5′ ribosomal binding sites (RBSs) and fused to a myc- or his-tag, respectively. The sequence was cloned downstream of the inducible pBAD promoter in the pBAD/HIS A plasmid and then transformed into BL21 to generate BL21-TAN. (B) BL21-TAN cultures were grown to an exponential phase (shaking at 37 °C) and then left alone (uninduced, U) or induced (I) at a final concentration of 0.02% L-arabinose for 4 h. Bacterial pellets and culture media (CM) were then subjected to western blot analysis to detect the expression of ColH (anti-Myc) and HylB (anti-His). Western blot original images can be found in Supplementary Materials. (C) BL21 transformed with a control pBAD-eGFP plasmid (BL21-eGFP), and BL21-TAN were cultured to an optical density (OD₆₀₀) of ~1. Cultures were either left uninduced (BL21-eGFP, BL21-TAN) or induced at 0.02% L-arabinose (BL21-TAN), and OD₆₀₀ was measured over time. (D) Uninduced or induced BL21-eGFP (0.02% L-arabinose, 4 h) were fixed in 4% paraformaldehyde and then stained simultaneously with anti-myc and anti-his to detect ColH and HylB expression, respectively. A representative single bacterium for each condition is shown. n.s. = no significance. Error bars display standard error of the mean.

Figure 2

In vitro degradation of collagens and hyaluronic acid by BL21-TAN. Hydrolysis reactions were performed using uninduced (U) or induced (I) BL21-TAN co-incubated with FITC-conjugated pig skin gelatin (A), bovine skin collagen type I (B), human placenta collagen type IV (C), or purified hyaluronic acid (HA) (D) in 50 mM Tris-HCl (pH 8.0) containing 10 mM CaCl₂ at 37 °C. The negative control includes the culture media (LB). Increases in fluorescence intensity signify the degradation of the FITC-conjugated target. Enzyme activity was measured by monitoring fluorescence (FITC) (ex: 495 nm, em: 519 nm). Data are expressed as mean ± error of the mean of three independent experiments. **** p < 0.0001, t-test.

Figure 3

BL21-TAN depletes PDAC-derived collagen and HA in serial tumor sections. Serial sections of KPC, BxPC3, and de-identified patient (UPN) PDAC tumors were treated overnight with BL21-TAN under uninduced or induced conditions at 37 °C. Sections were then stained by trichrome to detect collagen (blue). (A) or biotin-labeled hyaluronic acid binding protein (HABP) followed by streptavidin-FITC (B). Trichrome images were deconvoluted using ImageJ to quantify collagen content (blue staining) in randomly selected fields (10) of each tumor section (C). Fluorescence intensity was used to quantify HA content (FITC/488 channel) and was measured using ImageJ and normalized to uninduced treatment (D). Data are expressed as mean values ± error of the mean. ** p < 0.01; *** p < 0.001; **** p < 0.0001, t-test. Scale bar = 20 µm.

Figure 4

Intravenously administered BL21-TAN colonizes PDAC tumors and expresses both ColH and HylB. Mice bearing subcutaneous KPC or BxPC3 tumors (6–8 mm diameter) were intravenously injected with 5 × 10⁷ colony-forming units (CFUs) of BL21-TAN for three consecutive days. Twenty-four hours following the final injection, the mice were either administered PBS (uninduced, U) or 40 mg L-arabinose (induced, I) intraperitoneally. Tumors were collected 48 h post induction and sections were evaluated for BL21-TAN colonization and enzyme expression by immunofluorescence using antibodies specific to BL21 E. coli, Myc-tag (ColH) and His-tag (HylB). Objective: 100× oil. Scale bars = 10 µm.

Figure 5

In vivo collagen and HA depletion by BL21-TAN. NSG mice with subcutaneous BxPC3 xenografts (6–8 mm diameter) were intravenously injected with 5 × 10⁷ colony-forming units (CFUs) of the BL21-eGFP control or BL21-TAN for three consecutive days by the intravenous route. Twenty-four hours following the final injection, the mice given BL21-TAN were administered 40 mg L-arabinose intraperitoneally. Tumors (n = 8) were collected 48 h post induction and serial sections were evaluated for collagen (blue) using trichrome staining (A) and HA using HABP staining (B). Regions (box) from each image were magnified for greater resolution of collagen and HA (inset). Random fields (n = 15, 10× objective) from each treatment group were used for deconvolution analysis to quantify collagen and HA content in multiple tumors (C). Data are expressed as mean ± error of the mean. **** p < 0.0001, t-test. Scale bars = 1 mm.

Figure 6

BL21-TAN pre-treatment of PDAC tumors enhances the efficacy of gemcitabine. (A) NSG mice with subcutaneous BxPC3 xenografts (6–8 mm diameter) were injected with 5 × 10⁷ colony-forming units (CFUs) of the BL21-eGFP control or BL21-TAN for three consecutive days by the intravenous route. After 24 h of the final injection, mice were administered 40 mg L-arabinose (induction) or PBS control intraperitoneally. Gemcitabine (GEM) or vehicle (PBS) was injected intraperitoneally 48 h after induction and maintenance doses were given every three days thereafter. Tumor growth (B) and mouse weights (C) were measured over time until control groups required euthanization. Data are expressed as mean values ± error of the mean. *** p < 0.001; **** p < 0.0001, 2-way ANOVA.