GLuc A sr TK or AAV
GLuc-2A-sr39TK, or AAV2WT-CMV-GLuc-2A-sr39TK, respectively
(Fig. 1 C1, C2 and C3). There was high correlation of the images with serologic GLuc levels in AAV2RGD-BIRC5-SPTSTA-GLuc-2A-sr39TK, re- sulting in high serum GLuc levels in PDAC tumor bearing mice (Fig. 1 D1; black bar) and no levels seen in non-tumor mice (Fig. 1 D1; white
nograft PANC1 tumor model. Bioluminescence imaging of xenograft mouse model using systemic delivery of AAV2RGD-BIRC5-SPTSTA-GLuc-2A-sr39TK demonstrated a highly specific PDAC tumor signal (A1; red circle) with no background noise, which were quantified in B1 and C1,
respectively. Systemic delivery of control AAV2RGD-CMV-SPTSTA-GLuc-2A-sr39TK de- monstrated significantly less PDAC tumor signal (A2; red circle) with significantly greater non-specific background noise, quantified in B2 and C2, respectively. Systemic delivery of control AAV2-CMV-
SPTSTA-GLuc-2A-sr39TK demonstrated minimal tumor signal (A3; yellow circle) with even greater non-specific background
noise, quantified in B3 and C3, respectively. Imaging data of AAV2RGD-BIRC5-SPTSTA- GLuc-2A-sr39TK highly correlated with serologic GLuc levels (D1; black bar) vs absence of GLuc in control mice without tumor (D1; white bar). Imaging data of both control vectors did not correlate with ser-ologic GLuc levels (D2 and D3; black bars); nonspecific GLuc levels were found in con-trol mice without tumors (D2 and D3; white
bar). Immunofluorescence of tumors and benign tissues following systemic delivery of AAV2RGD-BIRC5-SPTSTA-GLuc-2A-sr39TK demonstrated viral thymidine kinase (vTK) was highly expressed only in PDAC tumors, but not in benign tissues (E). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
bar). Both control vectors resulted in serologic GLuc levels in both tumor bearing (Fig. 1 D2; black bar; D3; black bar) and non-tumor bearing mice (Fig. 1 D2; white bar; D3; white bar), demonstrating the lack of specificity to detect CFTRinh-172 of GLuc in PDAC versus benign tissues in mice. Mice were sacrificed and necropsy was performed; tu-
mors were excised and viral thymidine kinase (vTK) expression was measured in tumors and benign tissues. AAV2RGD-BIRC5-SPTSTA-GLuc-
2A-sr39TK resulted in high expression of sr39TK with no expression in benign tissues (Fig. 1, E). These data demonstrate the feasibility of a precision diagnostic platform utilizing AAV2RGD-BIRC5-SPTSTA-GLuc-2A-sr39TK to sensitively and specifically detect and localize minute human PDAC tumors in mice.
3.2. Selection of BIRC5 as a prototype target gene of PDAC in vitro
Studies, including our own, have shown that BIRC5 is overexpressed in the majority of PDAC [9–11]. Therefore, we selected BIRC5 as the prototype PDAC-upregulated gene for the platform.
To investigate BIRC5 expression in engineered models of PDAC, KrasG12D or/and truncated P53 were over-expressed in benign human primary pancreatic ductal epithelial (HPPE) cells. Western blot analysis
showed that BIRC5 expression was initiated in HPPE cells expressing KrasG12D and was augmented in HPPE expressing KrasG12D/truncated p53 (p53(1−320)) (Fig. 2A and B). Driver mutations, KrasG12D, p53 de-
letion (p53Del), p16Del, and SMAD4Del were introduced into HPPE cells using CRISPR/Cas9. Co-transfection with sgRNAs and Cas9 was in-dicated by the presence of green and red colors in HPPE cells, respec-tively (Fig. 2C); cells were further sorted and enriched by flow cyto-metry (Fig. 2D). Mutations were confirmed by Sanger sequencing,
successfully mutated. Upregulation of BIRC5 was evaluated using western blot analysis (Fig. 2E). KrasG12D alone resulted in low level
BIRC5 expression in engineered HPPE cells, with a robust increase in BIRC5 expression following the addition of p53Del mutation (Fig. 2E). However, there were no further increases in BIRC5 expression following the addition of p16Del and SMAD4Del (Fig. 2E). Therefore, we chose to focus on KrasG12D and p53Del mutations for the following studies. The time course for growth of state-of-the-art mouse 3D pancreatic ductal organoids with and without KrasG12D and p53Del driver muta- tions was captured using time-lapse video (IncuCyte S3, Essen Bioscience, Ann Arbor, Michigan) (Movie S1, S2). Co-transfection of lenti-gRNA (green) and lenti-Cas9tdt (red) was indicated within the organoids (Fig. 2F). Transformed organoids had tumor-like growth patterns (Fig. 2G, upper panel) with sgRNAs and KrasG12D-HDR (GFP)
and Cas9 (tdTomato), whereas wild-type organoids exhibited a circular pattern (Fig. 2G, bottom panel). Transformed organoids were observed within 5 days and displayed early PDAC morphology (Fig. 2F). Trans-formed organoid cross-sectional H&E staining, confirmed by 3 in-dependent pathologists, was consistent with PDAC morphology, whereas that of wild-type organoids resembled normal ductal struc-tures. Therefore, CRISPR/Cas9 engineering of pancreatic ductal orga-noids with driver mutations provides an excellent in vitro model to re-capitulate development of minute PDAC.