Early detection of pancreatic cancer in mouse models using a novel antibody, TAB004

Pancreatic ductal adenocarcinoma (PDA) is the fourth-leading cause of cancer death in the United States with a 5-year overall survival rate of 8% for all stages combined. But this decreases to 3% for the majority of patients that present with stage IV PDA at time of diagnosis. The lack of distinct early symptoms for PDA is one of the primary reasons for the late diagnosis. Common symptoms like weight loss, abdominal and back pains, and jaundice are often mistaken for symptoms of other issues and do not appear until the cancer has progressed to a late stage. Thus the development of novel imaging platforms for PDA is crucial for the early detection of the disease. MUC1 is a tumor-associated antigen (tMUC1) expressed on 80% of PDA. The goal of this study was to determine the targeting and detection capabilities of a tMUC1 specific antibody, TAB004. TAB004 antibody conjugated to a near infrared fluorescent probe was injected intraperitoneally into immune competent orthotopic and spontaneous models of PDA. Results show that fluorophore conjugated TAB004 specifically targets a) 1 week old small tumor in the pancreas in an orthotopic PDA model and b) very early pre-neoplastic lesions (PanIN lesions) that develop in the spontaneous PDA model before progression to adenocarcinoma. Thus, TAB004 is a promising antibody to deliver imaging agents directly to the pancreatic tumor microenvironment, significantly affecting early detection of PDA.


Introduction
Incidence and mortality trends predict pancreatic cancer will become the second-leading cause of cancer related deaths by 2020 in the United States. The mean expectation of life is less than six months and there are few long-term survivors. According to the Annual Cancer Statistics Review, patients with pancreatic carcinoma have the lowest 5-year survival rate [1,2]. Poor prognosis for patients is mainly due to late diagnosis, as a result of the lack of distinct early symptoms and effective diagnostics [3]. Only 15-18% of pancreatic cancer cases are resectable, and surgery offers the only single modality for potential cure. These patients have a two-year survival rate of 20%-40% with surgery, but despite surgical resection, local PLOS  recurrence or metastasis occurs in more than 50% of the patients (predominantly liver and peritoneum). Adjuvant therapy in patients with resectable pancreatic cancer including radiation and chemotherapy is a subject of controversy with randomized trials showing contradictory results. Very often the cancer becomes resistant to such therapies. Several of these therapies also produce undesirable side effects and in some instances damage to major organs. Overall survival from PDA is only possible with surgery and adjuvant treatment when detected early [4,5]. Thus, development of an effective and targeted detection platform is essential in order to improve the survival of PDA patients. Infiltrating PDA accounts for over 95% of all exocrine pancreatic malignancies. Activating mutations in the KRAS proto-oncogene are found in over 90% of invasive PDA and are thought to represent an initiating event. Recently a transgenic mouse model has been created that expresses physiological levels of oncogenic KRAS with a glycine to aspartate substitution at codon 12 (KRAS G12D ) in the progenitor cells of mouse pancreas. These mice develop the full spectrum of pancreatic ductal adenocarcinoma from preinvasive neoplasias (PanINs) to invasive and metastatic disease (designated as the Cre-LSL-KRAS G12D or PDA mice). Mucin-1 (MUC1), is a transmembrane protein with a heavily glycosylated extracellular domain [6]. Normal expression of MUC1 can be found on all glandular epithelial cells of the mammary gland, esophagus, stomach, duodenum, uterus, prostate, lung, and pancreas [7]. The negatively charged glycosylated extracellular domain of MUC1 in normal healthy tissues creates a physical barrier and an anti-adhesive surface, preventing pathogenic colonization [8]. However, in 80% of PDA, the extracellular domain of MUC1 is hypoglycosylated and the protein overexpressed [9]. This alteration of the structure and expression of MUC1 is associated with higher metastasis and poor prognosis [10,11] but also makes it the 2 nd most targetable tumor antigen [12].
Using the syngeneic KCM cell lines, we demonstrate that TAB004 specifically binds to tMUC1 expressing orthotopic KCM tumors in immunocompetent mice. Further, we show that TAB004 specifically targets the pancreas in the spontaneous tumor model (the KCM mice) at the early PanIN lesion stage much before the development of invasive PDA. We show that accumulation of TAB004 is significant at the tumor site but not so at other glandular epithelial organs. TAB004 can be further developed as a diagnostics imaging tool for early detection of PDA.

TAB004 production and comparison
TAB004 antibody (patent US 8518405 B2 and US 9090698 B2 provided by OncoTAb. Charlotte, NC) [16,[18][19][20][21][22]. Parental Murine TAB004 (mTAB004) is a mouse IgG1 monoclonal antibody created from a hybridoma. Chimeric TAB004 (cTAB004) is a chimeric IgG1 antibody with mouse ScFv and human Fc regions. cTAB004 was cloned into a Lake Pharma high expression stable cell line mammalian vector system (CHO cells). After two rounds of cell screening, three top performing CHO cells expressing the antibody were selected for production runs to assess the characteristics of the final stable cell line. The production of all versions of TAB-004 is currently conducted by LakePharma Inc., Belmont, CA. Comparison of mTAB004 and cTAB004 binding profiles were performed by ELISA. 100μl of 3μg/ml in 1x PBS of each antibody was used to coat 96-well ELISA plates (07-200-721, Fisher Scientific, Pittsburgh, PA) for 24 hours at 4˚C. Varying concentrations of KCM cell lysate were incubated for 1 hour at room temperature. Following incubation, detection antibody conjugated to HRP was added and incubated for 1 hour at room temperature. After 100μl of TMB (PI37574, Thermo Scientific, Waltham, MA) was allowed to incubate for 30 minutes, 50μl of Stop solution (SS03, Invitrogen, Waltham, MA) was added. All steps were followed by 5x washed with 1x wash buffer.

Confocal microscopy
Tumor/pancreas sections were treated with NucBlue Fixed Cell ReadyProbes Reagent (Ther-moFisher, Waltham, MA) for 5 minutes and Wheat Germ Agglutinin-488 (Molecular Probes, Waltham, MA) for 20 minutes. The slides where then washed with PBS for 5 minutes (3x) and fixed with 4% formaldehyde. Prolong Gold Antifade reagent with DAPI (P36935, Molecular Probes, Waltham, MA) was applied to mount coverslips. Images were acquired on an Olympus Fluoview FV 1000 confocal microscope.

In vivo experiments
C57Bl/6 mice were purchased from Jackson Laboratory and housed at UNC Charlotte's vivarium. For the orthotopic tumor model, C57/Bl6 female mice were injected in the pancreas with 5x10 5 KCM-Luc cells and allowed to recuperate for 7 days before any experiments were performed. For tumor and antibody visualization, orthotopic KCM-Luc tumor bearing C57/Bl6 mice were injected with 125μl of Redijet D-Luciferin (760504, Perkin Elmer, Waltham, MA) intraperitoneally and imaged 25 minutes later. For groups injected with antibody, 12.5μg of IgG1 Isotype conjugated to ICG or TAB004 conjugated to ICG was injected intraperitoneally and imaged at various time points with a Perkin Elmer IVIS Spectrum. KCM mice were generated in the Mukherjee lab. This mouse model is a triple transgenic cross of LSL-KRASG12D x P48-Cre x Human MUC1.Tg that will develop PDA spontaneously and express human MUC1 in a tissue specific manner [13]. The P48-Cre mice have a tamoxifen inducible promoter and therefore, oncogenesis is initiated only when mice are treated with tamoxifen (75 mg/kg in 100ul of corn oil, 1 injection per day for 2 weeks [10 days]. T5648-1G, Millipore Sigma, St. Louis, MO) as recommended by Jackson labs (https://www.jax.org/research-and-faculty/ tools/cre-repository/tamoxifen). Mice were euthanized at the end of all imaging studies. This study and all procedures were performed after approval from the Institutional Animal Care and Use Committee of UNC Charlotte.

Image analysis
All fluorescent slide images were analyzed using Image-J (National Institutes of Health, Bethesda, Maryland). All mice and organ images were processed in Living Imagine 4.3.1 (Caliper Life Sciences, Waltham, MA).

Immunohistochemistry
The pancreases of KCM mice were collected and samples fixed in buffered formalin and embedded in paraffin, and 5-to 6-μm-thick sections were obtained. Standard H&E staining protocol was performed and the tissue slides were then assessed by light microscopy to determine the PanIN lesions and progression to PDA at increasing disease stage. Microphotographs were taken using a DP70 camera and the Olympus Software Suite (Olympus, Waltham, MA)

Results
Targeting of tMUC1 in immunocompetent orthotopic PDA model with TAB004-ICG Control groups. Four control groups were imaged alongside TAB004 injected groups. Radiance efficiency (RE) values were collected using the Region of Interests (ROIs) for all images.
Control group 1 comprised of normal C57/Bl6 mice with nothing injected that had their organs imaged using the IVIS (S1 Fig). The purpose of Control group 1 was to determine background fluorescence (of ICG) levels in C57/Bl6 mice (Fig 1). Background fluorescence of the interior of the mice (in-situ) where the tumor is normally located (S1A Fig There was no significant increase in RE values between in-situ and organ images of Control group 1 and 2 (Fig 1), therefore background RE values do not increase if an orthotopic tumor is present.
Control group 3 and 4. Next, we were interested in seeing if a control IgG isotype antibody conjugated to ICG would accumulate in any regions of the mice. Control group 3 consisted of normal non-tumor bearing C57/Bl6 mice that were injected with IgG1-ICG (Fig 2). Exterior ICG images taken 24 hours post injection of the IgG1-ICG show no fluorescent signal stronger than the background of the control (Fig 2A). Interior ( Fig 2B) and organ ( Fig 2C)   values display no significant increase over values of Control group 1 or 2 (Fig 1). Control group 4 comprised of orthotopic KCM tumor-bearing C57/Bl6 mice injected with IgG1-ICG (Fig 3). Exterior ICG images taken 24 hours after injection of the IgG1-ICG showed no significant increase in fluorescence at the tumor site ( Fig 3A). Additionally, no significant increase in RE values from interior ( Fig 3B) and organ ( Fig 3C) images was observed between all the control groups (Fig 1). Therefore, the injection of an isotype control antibody with ICG does not increase background ICG RE values and any increase in ICG RE values seen in TAB004-ICG injected mice can be unequivocally taken as true accumulation of TAB004 at tumor region.
TAB004-ICG experimental groups. Murine monoclonal TAB004 antibody: The first treatment group consisted of KCM tumor-bearing C57/Bl6 mice injected with mTAB004-ICG (Figs 4 and 5). Exterior ICG images taken as early as 1 minute post injection (PI) show accumulation of the mTAB004 at the tumor site ( Fig 4A). The accumulation appears to increase 14 hours PI but decreases after 24 hours (Fig 4A). Exterior bioluminescent image of the same mouse confirm the location of the tumor (Fig 4B). In-situ photograph (Fig 5A) confirms the location of the tumor and IVIS images show significant accumulation of mTAB004-ICG at the tumor site (Fig 5B). The larger region of red-yellow fluorescence indicates accumulation of mTAB004 at the tumor (Fig 5B). Significant increases in RE values (Fig 5C) was observed in the pancreatic tumor, liver, small intestine, and spleen (Fig 1) which indicates the accumulation of mTAB004 selectively in those organs. RE values at the tumor site were~3.5 fold higher and~2 fold higher at the liver and small intestine site compared to the control groups (Fig 1). The increase in RE values in the liver is expected [25] from an IgG antibody. While the increase in RE values in the small intestines and spleen would suggest mTAB004 accumulation at the organs, and could be explained by some residual tumor in those organs that could not be completely dissected from the organ. Taken together, mTAB004 showed high specificity to the tumor and proves to be useful for detection of PDA by imaging.
Chimeric TAB004. Due to promising results with mTAB004, we generated a chimeric version of the antibody which shared an identical binding profile to mTAB004 in an ELISA  A shift to humanize the antibody is essential for further development of any antibody-based targeted imaging for diagnostics or for targeted therapy. We acquired a chimeric version of TAB004, cTAB004, which comprises of murine antigen recognition moiety in a human IgG1 backbone. KCM orthotopic tumor-bearing C57/Bl6 mice were injected with cTAB004-ICG (Figs 6 and 7). Similar to the mTAB004-ICG (Fig 4), exterior ICG images taken as early as 1 minute post injection show accumulation of the cTAB004 at the tumor site ( Fig  6A). The accumulation increases 14 hours PI but decreases post 24 hours (Fig 6A), similar to  mTAB004 injected mice. Exterior bioluminescent image of the same mouse confirm the location of the tumor (Fig 6B). In-situ photograph of a cTAB004 injected mouse confirms the location of the tumor (Fig 7A). IVIS in-situ image shows accumulation in the tumor and few other organs (Fig 7B). Dissected organs were imaged separately by IVIS to calculate the RE. The larger region of red-yellow fluorescence indicates accumulation of cTAB004 at the tumor ( Fig  7B). Significant increases in RE values were observed in the pancreatic tumor, liver, and small intestine (Fig 1); however, unlike mTAB004, there was minimal accumulation of the cTAB004 in the spleen. RE values at the tumor were~3.5 fold higher while~0.5 fold higher in the liver and~2.5 fold higher in the small intestine compared to control RE values (Fig 1). As with mTAB004, RE values in the liver is expected to be higher [25] with an IgG antibody; however, the liver was significantly lower in the cTAB004 compared to mTAB004 injected mice, possibly due higher uptake of chimeric antibody by macrophages [26]. The RE values of cTAB004 in the small intestines is the same as the mTAB004 treatment group, where residual secondary tumor bodies could not be removed from the organ.
All experiments shown in Figs 2-7 were conducted with n = 3 mice per experimental group and representative images from 1 mouse is shown for all groups. Fig 1 shows the ROI RE values plotted for n = 3 mice per group and significance are represented as p value.
Tumors from 24h post cTAB004-ICG injected mice were removed, fixed, paraffin embedded, sectioned (4micron sections) and imaged under a confocal microscope to determine the cellular localization of cTAB004 ICG (Fig 8). Tumors from 3 cTAB004-ICG mice are shown alongside a control tumor from mice that were not injected with cTAB004-ICG. Tumor sections from the cTAB004-ICG injected mice display significantly more ICG fluorescent signal (red) than control tumors without TAB004 ICG (Fig 8A) in all 3 mice. The blue and green fluorescent signals are DAPI (nucleus) and wheat germ agglutinin (membrane) respectively. Red ICG fluorescence was noted within the tumor bed as well as in the edges of the tumor. Most of staining seems to be localized in the surface and cytoplasm of the tumor cells. Green fluorescence suggests the disorganized membrane staining typical of undifferentiated tumors. Quantification of the red fluorescence showed significant increase in the corrected total cell fluorescence ( Fig 8B).
Overall, we have shown that cTAB004 behaves nearly identical to mTAB004 when binding to the pancreas/tumor site, with no loss in accumulation index, which was determined by the RE values (summarized in Fig 7). Expected accumulation in other organs, besides the liver, was not significantly higher than control groups. cTAB004 is highly specific for the tumor and provides rationale for further development of this platform technology for targeted imaging of PDA.

Targeting the pancreas before PDA develops in the KCM mice
Although we determined the specific targeting in an orthotopic tumor model, the question remains if this antibody can be developed for early detection of pre-neoplastic lesions prior to the development of PDA. As a starting point to address this question, we generated KCM mice by crossing the LSL-KRAS G12D with the tamoxifen inducible P48 Cre with the human MUC1. Tg mice [13]. These triple transgenic KCM mice develop spontaneous PDA when induced with tamoxifen. As controls, we generated KC mice that are double transgenic cross between LSL-KRAS G12D x tamoxifen-inducible P48 Cre mice. KC mice do not express the human MUC1. Tamoxifen was injected in KCM and KC mice to initiate oncogenesis while control KCM mice remained without tamoxifen (therefore no initiation of oncogenesis). The 3 groups of mice (tamoxifen-induced or un-induced KCM and KC mice) were injected with cTA-B004-ICG and imaged 24 hours PI (Fig 9). Both KC and KCM mice without tamoxifen induction did not display any ICG fluorescence signal in the pancreas from the exterior (Fig 9A and  9B left) or interior (Fig 9A and 9B right) IVIS images where the pancreas was imaged at its original and secondary positions. Secondary position represents moving the pancreas in situ using a forceps to confirm that any fluorescence signal is originating from the pancreas and nowhere else. As early as 3-weeks post tamoxifen induction, KCM mice showed ICG fluorescent signal over background from the exterior (Fig 9C left) and interior (Fig 9C right) IVIS images. Moving the pancreas from behind the liver and stomach displays a clearer ICG fluorescent signal over background (Fig 9C far right). Organs from these animals were also imaged with the pancreas possessing the highest level of RE over other organs, similar to the orthotopic tumor model (S3 Fig). None of the other organs showed fluorescence signal above background levels. It must be noted that in these KCM mice, all other glandular epithelial organs express normal human MUC1 but TAB004-ICG only accumulates in the pancreas post initiation of oncogenesis suggesting the high specificity of TAB004 to transformed/tumor associated form of MUC1. Based on our previous publication, we infer that at 3-weeks post completion of tamoxifen treatment, the mice have PanIN lesions 1a and b [13]. Additionally, KCM mice 11 weeks post tamoxifen induction, which develop PanIN 2 lesions at this point, were also injected with cTAB004-ICG to determine if it was possible to track disease progression ( Fig  9D). By this time, the pancreas is larger and the TAB004-ICG fluorescent signal shows accumulation at the pancreas as well. Moving the pancreas from its primary location provides a clear image of the TAB004-ICG signal in the pancreas (Fig 9D). The control tamoxifen uninduced KCM and tamoxifen-induced KC mice at the same age showed no TAB004-ICG fluorescence in the pancreas or any other organ (S4 Fig). The data clearly suggests that cTA-B004-ICG is binding to tMUC1at very early stages in PDA initiation (at the early PanIN stage) and is effective in tracking disease over time. All images are representative of n = 3 mice per group. Finally, H&E section of the pancreas 5, 8, 11, 22, and 33 weeks post tamoxifen injection confirms the formation of abnormal ducts at 3 weeks and PanIN lesions as early as 8 weeks (Fig 10).

Discussion
The concept of molecularly targeted diagnostic approaches would be very valuable toward the goal of precision medicine. The ability to monitor preneoplastic lesions and progression on a molecular level not defined by the presence of a palpable tumor mass and before any overt physiologic symptoms have developed would permit early and more adequate therapeutic intervention. This study addresses this need by examining the possibility of predictive diagnostics for PDA in a genetically engineered mouse model of pancreatic cancer. tMUC1 is overexpressed in over 80% of pancreatic cancer patients [27,28] and is also expressed in pancreatic intraepithelial neoplasia (PanIN) lesions, precursors of PDA [13]. This makes it a promising target for early detection and diagnostics of PDA. Specifically, we illustrate the detection of the changes in the expression profile of the molecular biomarker tMUC1 with a noninvasive imaging approach. By focusing on early disease, i.e., the PanIN lesion, we explore a scenario in which the molecular changes in tMUC1 precede changes in anatomical and physiologic signs of tumor development. This study extends our prior work in breast cancer, which demonstrated that changes in MUC1 antigen that occur in breast cancer development could detected in vivo using TAB004 as a carrier for imaging agent [17,29]. By broadening the application of Early detection and treatment of pancreatic cancer this methodology to pancreatic cancer, we move closer to establishing the value of tMUC1 as a wide-ranging cancer biomarker.
Progress has already been made in the direction of targeted cancer diagnostics. An example includes screening for the BRCA mutation for the assessment of breast cancer risk. With regard to noninvasive imaging, the development of dynamic MRI techniques, magnetic resonance spectroscopy and positron emission tomography have contributed to progress. Still, none of these technologies probes for specific molecular biomarkers expressed by cells in proportion to their potential for malignancy. Consequently, this highly specific molecular imaging approach has the potential for capturing the earliest signs of neoplastic transformation and in the future permit predictive diagnosis and response to therapy.
Our results from the orthotopic tumor model demonstrate that TAB004 is highly specific in targeting the PDA tumor and does not accumulate in other organs. In control group 1 (normal C57/Bl6 mice) and 2 (KCM-Luc tumor bearing mice-no TAB004 injected), we established background fluorescence levels for the mice and used these values to determine antibody accumulation at the tumor site and other organs. Control groups 3 (non-tumor bearing mice injected with isotype IgG1-ICG) and 4 (KCM-Luc tumor bearing mice injected with isotype IgG1-ICG) displayed no significant increase in fluorescence levels in all regions when compared to control groups 1 and 2, suggesting that from an imaging standpoint, non-targeting antibody at the chosen concentration clears from the mice in 24 hours. The significant increase in RE within the pancreas/tumor in mTAB004 and cTAB004 injected groups clearly demonstrates the retention of the targeting antibody. The livers from TAB004 injected groups displayed significant increase in RE over the control groups as well, with mTAB004 showing a greater increase over cTAB004. This appears to be a specific effect, possibly due to the presence of hepatic metastases from pancreatic cancer [30]. It also appears that TAB004 accumulates in the small intestine and spleen of mice, due to the increase in RE (Fig 7). Due to the orthotopic nature of the tumor, we have evidenced outgrowth of the primary pancreatic tumor into the surrounding organs including the spleen and small intestine due to their proximity to the pancreas. Whether this outgrowth is due to true metastasis or dissemination of tumor cells while injection is not clear at this time. Cells from the initial injection may leak out and cause many secondary tumors in the intraperitoneal space. Furthermore, we confirm that TAB004-ICG accumulates in the margins and within the pancreatic tumor bed (Fig 8). Other studies have attempted to target tMUC1 for diagnostic, imaging, and targeted therapy [17,[31][32][33]. However, due to the non-specificity of most of the MUC1 antibodies, we believe that TAB004 can improve the specific visualization of pancreatic tumor. Furthermore, fluorescent-tagged antibody may be useful in defining the tumor margins improving patient outcome that are eligible for resection [34].
Finally, we utilized a genetically engineered mouse model that spontaneously induced human tMUC1-positive pancreatic cancer (in KCM mouse) (Figs 9 and 10). The data from the KCM mice show that TAB004 can target tMUC1 being expressed in the pancreas before PDA develops. Examination of the pancreas from KCM mice show no evidence of a primary tumor, but IHC sections from these mice show the presence of PanIN 2 lesions. PanIN 2 lesions are the first stage in the development of PDA that is associated with significant genetic and molecular changes [35]. This early detection of the PanIN lesions can be translated into early detection of PDA and significantly improve disease outcome [5,36]. Recent studies have shown there is an association of pancreatic cancer with new onset diabetes [37][38][39]. In some studies, diabetes was determined to be present in nearly half of the pancreatic cancer patients at diagnosis, with 75%-88% of the cases of diabetes being new onset [40,41]. It is also interesting that other studies have shown that patients with new onset diabetes have a higher chance of developing pancreatic cancer [42][43][44]. Perhaps the methods used in this study can be adapted for early detection of pancreatic cancer in people with new onset diabetes.
The use of ICG as an imaging agent is a limitation in this proof of concept study demonstrating the ability of TAB004 to target early stages of PDA its tumors in an immune competent model. While ICG imaging is effective in mice, it may not be as effective in humans. As an advantage, ICG has a high photon count rate, but its depth of penetration is estimated to be between 2 and 4 cm [45], limiting its utility to imaging near the surface of the patients skin or patients undergoing surgery. There is a trend to shift from fluorescent imaging to radiolabeling targeting agents such as antibodies [46] to overcome this penetrance limitation. Single-photon emission computed tomography (SPECT) and positron emission tomography (PET) are the two majors molecular imaging modalities based on the detection of radioactive decay. Both PET and SPECT do not have limits in regards to their penetration depth and the image data is highly quantifiable [47][48][49]. For future studies we propose to use a fully humanized version of TAB004 with radioisotope labeling to target human xenograft tumors in immune compromised mice. Presently, this is outside the scope of this study.
As tMUC1 expression has global relevance in adenocarcinomas including pancreatic cancer, this study focused on applying our imaging approach to pancreatic cancer, as it has the lowest survival rate among all common cancers.