br Discussion br In this
In this feasibility study, a precision diagnostic platform was devel-oped to detect and localize minute human PDAC tumors in mice using an enhanced BIRC5 super-promoter to drive two reporter genes, de-livered by an engineered PDAC-specific adeno-associated virus. BIRC5, the gene that codes for the anti-apoptosis protein Survivin, has been defined as a target in many cancers, including PDAC, and was chosen as the prototype PDAC over-expressed gene . BIRC5 is prominently expressed during embryonal development, is absent in most terminally diﬀerentiated tissues, and upregulated in a variety of human cancers . It is over-expressed in 77–94% of PDAC, including pre-invasive PanIN3 and metastatic PDAC, but is absent in benign pancreatic tissue [9–11]. The organoid model and cell lines in this study revealed that BIRC5 is upregulated in the presence of Kras mutation and further augmented with P53 E-64-c confirming that common PDAC Kras and p53 driver mutations up-regulate expression of BIRC5 [20–23]. Inter-estingly, our PDAC modeling using pancreatic ductal organoids re-vealed tumor-like morphology, demonstrating that CRISPR/Cas9 en-
gineered organoids provide a powerful tool to recapitulate cancer development in real time [6,24,25]. Introduction of KrasG12D alone caused little change of morphology of ORGs in ten days with very mild
expression of BIRC5, which is consistent with transformations seen in the transgenic KC mouse model . However, introduction of KrasG12D and p53Del resulted in aggressive pancreatic ductal organoid morphological changes within five days, associated markedly increased BIRC5 expression. This transformation is consistent with the KPC mouse model, associated with aggressive pancreatic cancer and marked BIRC5 overexpression . These data, along with published data on BIRC5 over-expression in PDAC and most solid cancers, support the selection of BIRC5 as the prototype of PDAC over-expressed gene for the feasi-bility study.
To achieve translation, the diagnostic platform needs to be more sensitive and specific than existing diagnostic modalities for the de-tection and localization of minute PDAC tumors, including currently undetectable metastases. Traditional diagnostic modalities identify larger pancreatic lesions; the diagnostic sensitivity of endoscopic ul-trasound (EUS) is 95% for pancreatic tumors larger than 3 cm , whereas computed tomography (CT) has a sensitivity of 89%–97% with multiphasic helical detectors . Endoscopic retrograde cholangio-pancreatography (ERCP) and magnetic resonance cholangiopancreato-graphy (MRCP) have a sensitivity of 86–92% . However, each of these imaging modalities has poor sensitivity and specificity for minute PDAC tumors (< 10 mm) and benign precursor lesions, such as pan-creatic intraepithelial neoplasm (PanIN), mucinous cystic neoplasms (MCN) and intraductal papillary mucinous neoplasms (IPMN). EUS has proven superior for detection of early pancreatic neoplasms, as com-pared to non-invasive imaging-based diagnostic modalities, such as CT and MRCP . Only EUS with fine needle aspiration (EUS-FNA) is suﬃcient to definitively discriminate PDAC from other benign pan-creatic lesions. However, the diagnostic accuracy of EUS-FNA is limited by several factors: it is invasive, the mass needs to be localized near the GI tract, and requires availability of onsite cytopathologists and highly skilled endoscopists .
Radionuclide imaging using PET reporter probes are emerging as a valuable tool for detection of cancer in small animal models and are promising for clinical applications. HSV1-TK/[18F]FHBG, a noninvasive reporter system for PET imaging, is the focus of current imaging plat-forms . However, inability to deliver HSV1-TK specifically to PDAC tumors has been a significant limitation .
Clearly, there remains an urgent need for a precision diagnostic platform that can detect and localize minute PDAC and diﬀerentiate
resectable masses from currently undetectable metastases or benign pancreatic masses. To develop the platform, we employed a four part strategy: 1. selection of BIRC5 as a PDAC overexpressed gene for gen-eration of an enhanced BIRC5 super-promoter that would exponentially express two reporter genes; 2. selection of GLuc as a dual reporter gene, which is secreted by PDAC and therefore serologically detectable, as well as a localizing reporter in mice using optical imaging; 3. selection of a PET/CT reporter gene for localization of PDAC tumors in patients; and 4. engineering of a PDAC-specific viral delivery system. We then sought to determine whether the two reporter genes were suﬃciently expressed in minute human PDAC tumors in mice using two-step am-plification via an enhanced BIRC5 super-promoter coupled with the PDAC-specific AAV2 delivery system. As shown in Fig. 1, minute human PDAC tumors were sensitively and specifically detected and localized in mice with serologic Gaussia luciferase (GLuc) and optical imaging, re-spectively, thus demonstrating feasibility of the platform. Control vec-tors using a CMV promoter or an AAV2 virus lacking tumor homing peptides showed lack of specificity and considerable leakage in benign mouse tissues. There are recognized limitations of the feasibility study: