c-Src regulates cargo transit via the Golgi in pancreatic acinar cells

The exocrine pancreatic acinar cell is unique for its rapid protein synthesis and packaging in zymogen granules (ZGs). However, while crucial to the pathogenesis of pancreatitis, the signaling involved in the transit of proteins via the Golgi is poorly understood in these cells. Noting the evidence of c-Src in regulating transit of cargo via the Golgi in other systems, we explored this in acinar cells. Stimulation of ZG formation with dexamethasone activated Src and increased the Golgi area in acinar cells. c-Src localized to the microsomes of acinar cells on immunofluorescence and subcellular fractionation. While other Src family members had no effect on the Golgi markers P115 and GM130, active c-Src increased the Golgi area these stained, extending them into the ER. Src inhibition reduced amylase staining outside the Golgi and increased it in a stack like Golgi morphology. In vivo pharmacologic inhibition or acinar specific genetic deletion of c-Src reduced ZG number and staining of amylase in ZGs along with increasing amylase retention in the microsomal fraction. Morphologically this was associated with smaller Golgi stacks, and dilation of the endoplasmic reticulum. Therefore the role c-Src regulated Golgi function, ZG formation and microsomal zymogen transit in acinar cells needs to be explored in pancreatitis.


c-Src activation results in the Golgi expansion.
To study the effect of c-Src activation in vitro, we first adenovirally over expressed c-Src in primary pancreatic acinar cells. Low levels of Src expression ( Fig. 2A, cell2) resulted in an increase in P115 positive area compared to the untransfected cells (cells 3,4). This increase appeared in the cytoplasm/ER around unstained round areas resembling zymogen granules (inset yellow squares). This was also noted in AR42Jcells (Fig. 2B) where the Golgi marker GM130 behaved similarly. Interestingly, high level of Src expression (also seen in cell 1, Fig. 1A) diminished the Golgi marker staining intensity compared to un-transfected cells in which the Golgi remained compact (Fig. 2C left column). These effects of c-Src were dependent on its activation since pharmacologic Src inhibition with different inhibitors (Fig. 2C, SU6655; 10 μM, Dasatinib 10 μM, and PP2 10 μM) prevented its activation and the changes in Golgi morphology (Fig. 2B).
We next went on to study the Src family members that affected Golgi morphology. While Fyn, Yes, Lyn and c-Src could be overexpressed adenovirally (Fig. 3A), Yes was not spontaneously activated as noted by western blotting against the active form (pY416). Interestingly while all members showed a tropism for the Golgi area (Fig. 3B) and the cell membrane, only c-Src caused the morphological change in the Golgi (Fig. 3B,C), which were prevented by Dasatinib. Pharmacologic inhibition of Src reduces granule formation in AR42J cells. We then went on to study the role of Src in Golgi dynamics, along with trafficking of the exocrine enzyme amylase. We used dexamethasone (Dex; 1 μM) as previously 18 , which is known to increase amylase amounts. Dex caused Src activation (Fig. 4A). This was associated with an increase in Golgi area from 8.9 ± 3.5 µm 2 to 13.9 ± 5.9 µm 2 (p < 0.0001) which was reduced to 10.21 ± 3.6 µm 2 (p < 0.001) by Dasatinib (Fig. 4B,C).
To study the role of Src in the trafficking of amylase, AR42J cells were starved overnight. Next morning normal medium was replaced with alone Dexamethasone (Dex group) or after pretreatment with 10 micromolar Dasatinib (Dex + Das) and followed for another 24 hours. Dex resulted in a significant increase (*) in total amylase staining (Fig. 4D,E) most of which was round and granular (inset boxes in middle panel of Fig. 4D) along with being outside the Golgi (Fig. 4F). Dasatinib reduced the granular staining outside the Golgi ( † in Fig. 4F) along with increasing the amylase retained in the Golgi stacks ( † in Fig. 4G, and right most column in Fig. 4D, highlighted the yellow inset squares). It is to be noted that despite the close proximity of Golgi and granules in the Dex group, majority of the granules did not co-localize with the Golgi (Green inset squares) or were outside the Golgi (red inset squares).  To specifically study the role of c-Src in zymogen granule formation, we genetically deleted it in adult (12-16 week mice) by treating c-Src L/L /Cre positive mice with tamoxifen ( Fig. 6A). c-Src deletion did not affect the amount of amylase or chymotrypsin protein compared to the vehicle of tamoxifen (corn oil), but resulted in a decrease in the pink staining zymogen granule area in exocrine acinar cells from 46.8 ± 7.1% to 29 ± 4.6% (p < 0.01; Fig. 6B,C) similar to the Dasatinib treated mice. On immunofluorescence staining, this was seen as amylase having a diffuse cytosolic staining vs. the normal dense granule staining seen in mice without c-Src knock down (Fig. 6E) or wild type mice (Fig. 5C, upper row). c-Src knockdown also resulted in smaller Golgi stacks ( Fig. 6D) with morphologically less branching ( Fig. 6E lower panel, right side).
On subcellular fractionation, c-Src knockdown resulted in loss of enrichment of amylase in the zymogen granule fraction normally seen without c-Src knockdown, such as in the corn oil treated group (Zymo.; Fig. 7B, vs. 7A). The knock down caused a concurrent increase of amylase in the microsomal fraction (S1 to 10 5 gP) where the ER and Golgi markers Golgin-84 and Calnexin are present. This data correlates well with the loss of amylase staining in zymogen granules induced by tamoxifen (Fig. 6E). On electron microscopy ( Fig. 7C,D), the c-Src knockdown mice has a decrease in zymogen granule number, and an increase in the proportion of small size granules, along with the ER dilation. These findings support the role of c-Src in trafficking of amylase from the Golgi and ER into zymogen granules and in the formation of zymogen granules in acinar cells.

Discussion
In this study we investigated the role of c-Src on Golgi function and the trafficking of amylase, an exocrine enzyme secreted by pancreatic acinar cells. To do so we first identified the c-Src as a family member that is present on the microsomes of acinar cells. While other Src family members, i.e. Yes, Lyn and Fyn could be overexpressed, these did not perturb Golgi morphology. C-Src overexpression however resulted in morphological expansion of the Golgi markers GM-130 and P115 into the ER. This phenomenon was dependent on c-Src activation and was prevented by inhibiting it pharmacologically. We go on to show that stimulation of ZG formation in AR42J cells is associated with Src activation and expansion of the Golgi area also, and that pharmacologic inhibition of Src results in a reduction of amylase outside the Golgi and an increase in its retention in the stack like morphology of the Golgi. Similarly in vivo both pharmacologic inhibition of Src and its selective ablation in acinar cells of adult mice caused retention of amylase in the microsomes, dilation of the ER and a reduction in ZG number. Previous studies have shown Src family members to regulate upstream signaling in acinar cells including store operated calcium entry 17 , protein kinase c-delta activation via Lyn 20 , physiologic phenomena including endocytosis 12 , exocytosis 13 and pathological phenomena including blebbing 15 , trypsinogen activation 8 and chemokine upregulation 21 . In most of these studies however the Src family member responsible for the biologic phenomenon has remained unclear since the approach has traditionally been based on pharmacologic inhibition, which affects all Src family members. The current study provides the first specific role of a Src family member, i.e. c-Src in pancreatic acinar protein trafficking.
Src Kinases have been shown to reside on the Golgi 7,22 , and Src has been shown to be activated via transit of secreted cargo from the ER to the Golgi 11 . Src activation has also been shown to result from ER lumen proteins with the KDEL sequence reaching the cis-Golgi and binding the KDEL receptor, initiating a COP-1 sequence 23 or intra-Golgi trafficking 10 . While some studies show Src activation to result in fragmentation of the Golgi 11 , others show a redistribution of N-galactosylaminotranferases from the Golgi to the ER 9 regulating O-glycosylation, without affecting cis-Golgi markers such as GM130. Here we note Src to affect both Golgi dynamics and cargo transit-i.e. c-Src activity induced fragmentation of the Golgi and Src inhibition caused accumulation of amylase (a normally secreted protein without a KDEL sequence) in the ER, Golgi-with resultant ER dilation. The mechanism of this remains unknown. Previous studies inhibiting O-glycosylation showed a reduction of ZG formation in acinar cells 24 due to impaired condensation of secreted proteins in the TGN. Whether this kind of a phenomenon or a Src regulated cytoskeletal or budding event on the Golgi results in the finding we note remains to be studied. Knowing these can impact both the physiologic modulation of acinar function such as during fasting and refeeding, along with helping us understand the therapeutic relevance of targeting Src during pancreatitis.

Methods
Animals and animal procedures. C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, Maine) and used at 12-16 week age. These were housed with a 12-h light/dark cycle, at temperatures from 21-25 °C, fed standard laboratory chow, and allowed to drink ad libitum. Animals were acclimatized for at least 2 days before use. For genetic deletion of c-Src in the exocrine pancreas, mice with LoxP sites located 5′ of intron 4 and intron 7 of c-Src (NCBI NM_009271) with an FVB/NJ background 25 were backcrossed to a C57BL/6 background over 5 generations using the 150-180 SNP panel selected from publically available databases. This was done at the Jax Genome Scanning Service (Jackson Laboratory Bar Harbor, Maine). Dual  floxed (c-Src L/L mice) were identified using the primers (Forward: GGTCTTGTCATGGCTCTGTC, reverse: CATCTCTGCTCACCTGATAG) yielding a single 450 BP band, wild types yielding a 400 BP band as described previously 25 . c-Src L/L mice were then bred to mice expressing the fusion product of Cre recombinase and a mutant human estrogen receptor ligand binding domain (CreERT2) under the elastase1 promoter [Tg(Ela1-Cre/ ERT2)1Stof/J, Jackson labs, Stock # 08861], with the mice expressing Cre recombinase yielding a 500 BP band using CTCTGCTAACCA TGTTCATGCCT as forward primer and ACG CTAGAGCCTGTTTTG as reverse primer. Genetic deletion of c-Src was achieved by administration of tamoxifen (150 mg/kg, in 0.1 ml corn oil, intraperitoneal) to c-Src L/L Cre positive mice as 7 injections over 14 days followed by a 1 week recovery period. This was optimized based on 5 injections with 5 day recovery period at the same dose being ineffective in knocking down the c-Src protein. The control group received only corn oil. At the end of the recovery period, the mice were euthanized using carbon dioxide and the pancreas was immediately procured and used for histological, morphological, biochemical assays as described below. Dasatinib (50 mg/kg/day × 2 days) was given intraperitoneally as described previously 26 . In vivo studies are representative of 6-8 animals in each group. Cell lines and pancreatic acini. Primary pancreatic acini were harvested as described previously 15 from mice and cultured in RPMI 1640 with 10% Fetal bovine serum (FBS; GE life sciences, Logan, UT) at 37C in a 5% CO 2 -humidified incubator 27 alone or with adenoviruses. Viability next morning before use was more than 95% by trypan blue exclusion. AR42J cells (American Type Culture Collection no. CRL-1492) were cultured in F-12K in 10% FBS with incubator settings as in primary acini. These were used for studies on ZG formation since primary acinar cells do not make ZG in culture. Dexamethasone (1 μM) was used to induce protein synthesis and zymogen granule formation as described previously 28  Western Blotting. Proteins were extracted from pancreas or cells after homogenization with a Potter-Elvehjem homogenizer in homogenization buffer containing 50 mM Tris at pH 7.2, 150 mM NaCl, 0.5 mM EDTA, 1 mM EGTA, 2 mM dithiothreitol, 1 mM Na 3 VO 4 , 25 mM NaF, 1% NP-40, and Complete (Roche Diagnostics, Indianapolis, IN) protease inhibitor cocktail. Immunoprecipitation of Src and blotting for active Src was done as described previously 15 . For blotting, lysates were boiled in 1x Laemmli sample buffer before Western blot analysis according to standard procedures as previously described 15 .
Subcellular fractionation. This was done on whole pancreatic homogenates as described previously 6 to yield the zymogen and mitochondrial fractions. The mitochondrial fraction supernatant was further fractionated into the Golgi (S1, S2) and ER fractions (S3, Pellet) as described in the protocol by Taylor et al. 29 . Proteins were quantified in these and adjusted to a final concentration of 0.1 μg/μl, boiled in Laemmli buffer for western blotting. The full length un-cropped images are presented in the Supplementary Information. Immunofluorescence Microscopy. Immunofluorescence microscopy was done on pancreatic tissue cryosections embedded in Tissue-Tek® Optimal Cutting Temperature (OCT) (Sakura Finetek USA, Inc., Torrence, CA), AR42J cells or acini plated on plain glass coverslips. These were fixed with 2% paraformaldehyde and processed as described previously 15,30 . After blocking with 5% normal goat serum, tissue cryosections were incubated with primary antibodies (1:50 for cryosections and 1:200 for cells) for 1 h at room temperature, washed, and incubated with secondary antibodies (Alexa 488-or Alexa 594-conjugated, diluted 1:500) Cy5-conjugated phalloidin (100 nM) (ThermoFisher, Waltham, MA) with or without DRAQ5 (1:5000) for 30 min. After washing and mounting in Fluoromount G (Sigma Aldrich) on SuperFrost slides (ThermoFisher), confocal imaging (1 µm thick) was done using a Zeiss LSM800 confocal microscope (Thornwood, NJ). Images were collected with a Zeiss C-Apochromat 63x/1.2 NA water immersion objective.
Quantitative morphometry. Images were acquired and saved as. czi files. Golgi area, total amylase area, and amount of amylase out of the Golgi were quantified using automated image cytometry performed with FIJI/ImageJ (www.fiji.sc) using a custom macro. Briefly, images collected with Golgi in the first channel and Amylase in the second channel were collapsed as maximum intensity projections. Split channels were automatically contrasted with a saturation factor of 0.2 and converted to binary masks. Golgi and Amylase masks were auto-selected by white/black discrimination and area in µm 2 calculated according to the meta-info embedded scaling factor. Percentage of Amylase outside of the Golgi was calculated by subtraction of the Golgi mask from the amylase mask divided by the total area of the Amylase mask. 40-45 cells were counted for each measurement. The number of Golgi stacks with amylase in a stack like morphology was counted manually for each cell. Zymogen granule area in hematoxylin and eosin stained paraffin sections was quantified by measuring the area SCientifiC REPORTS | (2018) 8:11903 | DOI:10.1038/s41598-018-30370-4 occupied by zymogen granules and divided by the total cell area to give a % zymogen area. 7-10 acini were measured for each field captured on a 20x objective in a blinded fashion. Images were converted to. tiff format organized and processed using Adobe Photoshop CC (Adobe Systems, Mountain View, CA) for depicting in figures.

Statistical Analysis.
All values are presented as mean ± SEM. At least 3 independent experiments done for each data presented from in vitro studies. Differences between 2 groups were analyzed by unpaired Student's t test while ANOVA (Dunnett's method) was carried out to make comparison between continuous data from multiple groups. P value of <0.05 was considered to indicate statistical significance.
Ethics statement. All animal experiments and relevant methods were approved by the Institutional Animal Care and Use Committee of the Mayo Clinic foundation.