The MIT and UC Davis team engineered the probiotic bacterium E. Coli Nissle 1917 (EcN) with different stabilized expression cassettes. One, a PROP-Z (programmable probiotics with LacZ) construct, produces: i) a luminescent signal for whole-body imaging that is distinct from the signal produced by luciferase-tagged tumor cell lines, and ii) β-galactosidase, which can be detected enzymatically in urine. Another construct, PROP-Luc, only produces the luminescent signal for whole-body imaging. Bacteria engineered to carry these constructs can be delivered orally, and can detect tumors in liver.
BALB/c mice engrafted with the luciferase-expressing mouse colorectal tumor cell MC26-LucF delivered directly into the spleen developed metastatic liver tumors. Tumor growth was monitored noninvasively by MRI and using an in vivo imaging system (IVIS). Once tumors reached an average of 2-3 mm in diameter (3-5 weeks following engraftment), the mice were inoculated orally with PROP-Luc-modified EcN. One day following administration, 88% of the metastatic liver tumors displayed PROP-Luc-dependent luminescence. Tumor size ranged from 1 to 8 mm in diameter. Oral administration of PROP-Z-modified EcN resulted in similarly high liver tumor colonization levels. Seven days following PROP-Z EcN administration, unengrafted control mice showed no bacterial colonization based on qPCR, colony counts, and histological examination of various organs. The PROP-Z EcN-treated mice lived for at least a year and displayed no detrimental health effects. In another control experiment, FVB/NJ (001800) mice fed chow containing DDC (3,5-diethoxycarbonyl-1,4-dihydrocollidine) for three weeks developed liver fibrosis. Orally-delivered PROP-Z EcN failed to colonize the fibrotic tissue in these mice, confirming that the programmable probiotics specifically colonize liver tumors, as opposed to damaged liver tissue, and are tolerated well in mice.
The MIT/UC-Davis team next tested whether their engineered probiotic EcN could colonize tumors with different tissue origins and in other genetic backgrounds. They subcutaneously implanted human ovarian, colorectal, and pancreatic tumors in nude mice and mouse colorectal (MC26-LucF) and lung (393M1-LucF) tumor cells in BALB/c and B6129SF1/J (101043) mice, respectively. A single intravenous (IV) of PROP-Z-modified EcN at a high dose into the tumor-engrafted mice resulted in 100% colonization of all the subcutaneous tumors, except for mice that had been engrafted with 393M1-LucF lung tumors, in which only 83% of the injected mice showed EcN-colonized tumors. PROP-Z EcN delivered orally to the MC26-LucF- and 393M1-LucF-engrafted mice, however, failed to colonize the subcutaneous tumors, even at high doses. Colony counts following oral delivery failed to detect any bacteria in the blood, suggesting that the probiotic bacteria do not enter the circulation when administered orally, and this likely accounts for their inability to colonize subcutaneous tumors. These results demonstrate that the proclivity of PROP to colonize tumors is dependent on the ability to deliver the bacteria to the organs, not on the tumor type or origin.
Oral delivery of PROP-Z-engineered or non-LacZ control bacteria to mice with MC26-LucF liver metastases resulted in colonized tumors. Intravenous injection of LuGal, a soluble conjugate of luciferin and galactose that is converted to luciferin by β-galactosidase, to the engrafted and EcN-colonized mice produced easily detectable luciferase signal in the urine from mice colonized with PRO-Z EcN, but not from mice colonized with the non-LacZ control bacteria. No luciferase signal was detected in the urine following PROP-Z administration to control mice that did not carry liver metastatic tumors. Intravenous administration of CPRG (chlorophenol red‐β‐D‐galactopyranoside), a galactoside analog, to mice with PROP-Z colonized tumors changed the color of the animals’ urine compared to control PROP-Z-treated mice with no tumors.
These results demonstrate that this programmable probiotic platform can readily and specifically diagnose mice that carry liver metastases noninvasively by whole body imaging or via simple urine tests. These data also demonstrate the potential of synthetic biology as diagnostic and therapeutic tools.