br the hypothesis that GLuc A sr TK
the hypothesis that GLuc-2A-sr39TK can be used as a precision diagnostic platform for minute human PDAC. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
localization of PDAC, we performed transient transfection of PDAC cells with a CMV-driven GLuc-2A-sr39TK vector, which resulted in similar levels of secreted GLuc to that of CMV-GLuc (Fig. 4B) and equivalent expression of sr39TK to that of CMV-sr39TK (Fig. 4C). sr39TK activity was not aﬀected by the expression of GLuc-2A-sr39T (Fig. 4D). These data demonstrate eﬀective cleavage and separate expression of GLuc and sr39TK in vitro.
Human PDAC xenograft mouse models were generated using en-gineered patient-derived PDAC cell line-15, (PDCL-15CMV−GLuc−2A-sr39TK), Mia PaCa2CMV−GLuc−2A-sr39TK, or non-transfected Mia PaCa2
cells (6 mice/group: 3 males/3 females). Serologic GLuc levels were detected at 10 mm3 (1 mm diameter) PDAC tumor volume, whereas GLuc levels were undetectable in control mice (Fig. 5A). Rising ser-ologic GLuc levels highly correlated with increasing PDAC tumor vo-lume (Fig. 5B). GLuc Midostaurin (PKC412) imaging signals were highly pro-portional to tumor volume (R = 0.92; p = 0.005) (Fig. 5C and D). These data demonstrated remarkable sensitivity of serologic and optical imaging GLuc for detection and localization of minute human PDAC in vivo.
To evaluate microPET/CT imaging of human PDAC tumors, PDCL-15CMV−GLuc−2A-sr39TK cells versus control PDCL-15WT cells were im- planted in the right and left flanks, respectively, of the same mouse. When tumors reached 10 mm diameter, defined as a minute human PDAC tumor, mice underwent [18F]FHBG microPET/CT (PDAC-specific imaging) and [18F]FDG microPET/CT imaging (standard, non-specific imaging), as well as bioluminescence imaging. Three hours after in-jection of the specific TK probe, [18F]FHBG, microPET/CT revealed highly specific imaging of minute PDAC tumors only in the right flank
(Fig. 5E), which was consistent with results seen with bioluminescence imaging following injection of coelenterazine (Fig. 5F). Conversely, injection of [18F]FDG microPET/CT, resulted in non-specific signals seen in both tumors and normal tissues (Fig. 5G). Therefore, expression of GLuc-2A-sr39TK in human PDAC tumors was sensitively and speci-fically detected by microPET/CT imaging, which correlated with op-tical imaging. These data support the hypothesis that PDAC expressing dual reporter genes could be used to detect and localize minute human PDAC tumors in mice using serologic and optical imaging GLuc and sr39TK/microPET/CT imaging, respectively.
3.5. Genetically-engineered recombinant adeno-associated virus 2 (AAV2RGD) enhances specific gene delivery in PDAC
AAV2 was chosen as the delivery system due to its high gene de-livery eﬃciency and proven safety in clinical trials [16,17]. AAV2 uses cell surface heparan sulfate (HS), a highly sulfated polysaccharide, as a receptor to establish infections. However, PDAC cells lose HS expres-sion on their surface due to high levels of heparinase-1 (HPA-1) ex-pression (S8), which results in low infection by wild type AAV2. Therefore, we engineered the fundamental cellular attachment receptor of AAV2, heparin sulfate proteoglycan (HSPG), with tumor homing peptides (TumorHoPe) on the viral capsid R588 site to enhance gene delivery specifically to PDAC (S9). Eighteen TumorHoPe sequences were selected from the TumorHoPe database (http://crdd.osdd.net/ raghava/tumorhope/) (sTable2), which contains 744 entries of ex-perimentally characterized TumorHoPe. Capsid mutations at R585A and R588A served as controls. PDAC-specific AAV2 viruses were
Fig. 6. Using a screening assay, five tumor homing peptide-engineered AAV2 viruses (AAV2TumorHoPe) with the highest infectivity
were selected for validation in PDAC cells. GFP assay demonstrated that AAV2TumorHoPe
enhanced viral infectivity specifically in PDAC cells, but not in benign HPDE cells (6A; top panel). AAV2WT had far less viral infectivity in PDAC cells, and marked in-fectivity in benign HPDE cells (6A; bottom
panel) (scale bar = 100 μm). GLuc reporter assay demonstrated that AAV2TumorHoPe en-
hanced viral infectivity specifically in PDAC cells, but not in benign HPDE cells (6B; red bars). AAV2WT had significantly less viral infectivity in PDAC cells, and equivalent infectivity in benign HPDE cells (6B; blue bars). The data demonstrate high specificity of AAV2TumorHoPe for PDAC delivery of re-porter genes. (For interpretation of the re-ferences to color in this figure legend, the reader is referred to the Web version of this article.)
identified using a CMV-driven firefly luciferase (CMV-FLuc) reporter assay in PDAC cell lines versus benign HPDE and HPPE cells (S10). Optimal vectors in each cell line were selected based upon the highest infection rate of PDAC cells and the lowest infection rate of benign human pancreatic cells. AAV2RGD optimized targeting in PANC-1, Capan2, AsPC1 with no activity in benign HPDE and HPPE cells, therefore these engineered viruses were selected for further validation using GFP and GLuc reporter assays in PDAC cells. GFP expression (Fig. 6A) and GLuc expression (Fig. 6B) in each matched cell line were consistent with the findings of the FLuc reporter assay (S13). To further determine whether the specificity of infection was dependent upon specific ligand receptor binding, a competition assay was performed and revealed that high dose peptides inhibited infection of each cell by its corresponding virus (S11). Since integrin was highly expressed in the PDAC cells, but not benign human pancreatic cells (S12) , AAV2RGD was selected as the delivery system and was shown to be highly specific for PDAC cell delivery as seen in Fig. 1.