Loss of PACS-2 delays regeneration in DSS-induced colitis but does not affect the ApcMin model of colorectal cancer

PACS-2 is a multifunctional sorting protein that mediates cell homeostasis. We recently identified PACS-2 in a functional genome-wide siRNA screen for novel regulators of the metalloproteinase ADAM17, the main sheddase for ligands of the ErbB receptor family. Of note, we showed that Pacs2-/- mice have significantly reduced EGFR activity and proliferative index in the intestinal epithelium. As EGFR signaling is highly mitogenic for intestinal epithelial stem cells, and plays essential roles in intestinal epithelial regeneration and tumor development, we have now examined the role of PACS-2 in these processes. Specifically, we analyzed the role of Pacs2-deficiency in a DSS-induced colitis model as well as in the genetic ApcMin colon cancer model. We now report that loss of PACS-2 delays tissue regeneration after colonic injury with little effect on key inflammatory parameters. We did however not observe any apparent effects on tumor formation driven by excessive proliferative signaling downstream from APC-deficiency. Our findings reveal that the role of PACS-2 in regulating ADAM17-mediated shedding is not an obligate requirement for the epithelium to respond to the strong inflammatory or tumorigenic inducers in the models assessed here.


INTRODUCTION
Phosphofurin acidic cluster sorting protein 2 (PACS-2) is a multifunctional sorting protein involved in trafficking of various cargo molecules. PACS-2 frequently acts in a complimentary manner to its close relative PACS-1, and the PACS protein family is central to ensuring the correct sorting and localization of a multitude of functionally diverse client proteins, with numerous downstream biological effects [1,2]. We recently identified PACS-2 in a functional genome-wide siRNA screen for novel regulators of the transmembrane protease, a disintegrin and metalloproteinase-17 (ADAM17) [3]. We demonstrated that PACS-2 knockdown decreased the recycling of internalized ADAM17 protein, with a consequent reduction in the amount of active ADAM17 on the cell surface [3].
The dynamics of cell-surface localized ADAM17 are of importance because it cleaves a variety of cellsurface molecules, a process termed ectodomain shedding. Notably, ADAM17 sheds ErbB ligands, leading to activation of ErbB family receptors (epidermal growth factor receptor (EGFR), ErbB2, ErbB3, ErbB4) to trigger cell growth, survival, and migration [4]. ADAM17 is thus an important player in the regulation of epithelial development, growth and tissue homeostasis, and its deregulation has been implicated in many human diseases [5]. By decreasing cell-surface levels of ADAM17, PACS-2 knockdown led to a significant reduction in shedding of EGFR ligands. In vivo, this translated into reduced EGFR activition and proliferative index in the small intestinal epithelium and to a lesser extent in the colon of Pacs2 -/mice [3]. By controlling ADAM17 cell-surface availability, PACS-2 thus appears to fine-tune ADAM17

Priority Research Paper
and ErbB activity, with potential functional effects on intestinal homeostasis.
Intestinal epithelial homeostasis relies on stem cells in the mucosal crypts, which continuously replenish cells lost from the apex of villi [6]. EGFR signaling is highly mitogenic for intestinal stem cells [7] and therefore central for intestinal homeostasis and regeneration. Importantly, ADAM17-mediated EGFR activation is essential for gut regeneration, and Adam17 and Egfrdeficient mice are more susceptible to colitis induced by oral administration of the sulfated polysaccharide dextran sodium sulfate (DSS) [8][9][10][11]. Ingestion of DSS ablates the colonic epithelium, resulting in inflammation and subsequent repair of the damaged areas [12]. While the precise mechanisms governing ADAM17-and EGFR-mediated protection against DSS-induced colitis remain unknown, the ErbB ligands and known ADAM17 substrates amphiregulin (AREG), epiregulin (EREG), and transforming growth factor (TGF)-α appear to be key effectors [8,9,[13][14][15].
We hypothesized that as a regulator of ADAM17 and EGFR activities, loss of PACS-2 would influence intestinal epithelial pathologies. To this end, we treated Pacs2 -/mice and wildtype littermates with DSS and evaluated disease progression and effects on intestinal histology. Another process that relies on EGFR signaling is the development of intestinal tumors [16], which originate from crypt stem cells [17]. Accordingly, we investigated the effects of PACS-2 loss in the Apc Min mouse model of colorectal cancer. Apc Min mice mimic familial colorectal cancer in patients with only one functional adenomatous polyposis coli (APC) gene allele, and APC is furthermore mutated in most sporadic human colorectal cancers [18]. As in humans, loss of heterozygosity of the remaining wildtype Apc allele results in the formation of multiple adenomas, and the Apc Min/+ mouse is therefore a commonly used model of colorectal cancer [19].

Pacs2 -/mice show no significant changes in susceptibility to DSS-induced colitis
To evaluate the effects of PACS-2 on intestinal inflammation and regeneration, groups of male Pacs2 -/and littermate Pacs2 +/+ (control) mice were administered 2.3% DSS in the drinking water, followed by 5 days on regular water ( Figure 1A). Quantitative RT-PCR of mRNA isolated from colonic tissue of control and Pacs2 -/mice verified the Pacs2-deficiency of Pacs2 -/mice and importantly, revealed a small but statistically significant increase in the mRNA expression of PACS-1, the other member of the PACS protein family ( Figure 1B). Of note, Pacs2 mRNA expression was not significantly altered during DSS-induced colitis ( Figure 1C). During the course of colitis, Pacs2 -/mice tended towards poorer survival than littermate controls (35 % vs. 61 % survival at day 10, respectively; Figure 1D); however, this difference was not statistically significant (p = 0.08). Colon shortening, a macroscopic parameter of colitis severity, was observed to a similar extend in Pacs2 -/and control mice ( Figure 1E). Additionally, we found no significant difference in weight loss in Pacs2 -/mice as compared to controls ( Figure 1F).
To assess disease severity in more detail, we calculated the Disease Activity Index (DAI), adapted from the method first described by Cooper et al. [12]. The DAI is the average of three independent scores describing weight loss, stool consistency and stool blood content. The scoring system employed is outlined in Table 1. Although computation of the DAI showed a slightly increased disease severity in Pacs2 -/mice as compared to controls, the change was not statistically significant ( Figure 1G). Thus, while there was a tendency toward worse disease progression in Pacs2 -/mice undergoing DSS-induced colitis, these changes were not significant.

Pacs2 -/mice show comparable inflammation and tissue damage during DSS-induced colitis, but a slight delay in regeneration
Histological examination revealed severe tissue damage with numerous mononuclear cells infiltrating the lamina propria of colons on day 6 of the DSS-induced colitis model in both control and Pacs2 -/mice ( Figure  2A). Day 6 was the time of most severe tissue damage, with some tissue regeneration being seen in histological sections from day 10 ( Figure 2A). Tissue damage was quantified by a method adapted from Obermeier et al. [20]. Histological sections were scored blindly using the criteria outlined in Table 2 and a histology score was calculated by combining the epithelial damage score and the inflammation score. Consistent with the hematoxylin and eosin (H&E) images presented, there was no difference in the histology score between groups on day 6 of DSSinduced colitis ( Figure 2B, left graph). On day 10, the histology score was significantly higher in Pacs2 -/mice as compared to control mice ( Figure 2B, right graph), which indicates a slight delay in tissue regeneration.
To examine the regenerative response of the intestinal epithelium to DSS treatment, we assessed the level of activated EGFR (Tyr1068 phosphorylated) by immunofluorescence staining of histological colon sections from Pacs2 -/and control mice ( Figure 3A and 3B). While no difference was observed at day 6 after DSS treatment, the level of activated EGFR was significantly reduced at day 10 of the DSS-induced colitis model in Pacs2 -/mice compared to control mice. Yet, proliferation of the intestinal epithelial cells was not significantly affected at any of the time points, as assessed by Ki67 Disease Activity Index = Average of the sum of the three scores. The scoring system is adapted from that described by Cooper et al. [12].  Table 2. A. Representative images of colonic specimens from control and Pacs2 -/mice isolated at day 6 and day 10. Scale bar overview pictures = 1 mm. Scale bar detail figures = 0.1 mm. B. Dot plots show the histology score, calculated as the sum of the epithelial damage score and the inflammation score. The individual scores were calculated as described in Table 2. All data represent the mean ± S.E.M. and were analyzed by unpaired Student's t-test (n = 6-13 as indicated). Data were pooled from three independent experiments. staining ( Figure 3C and 3D). Since PACS-2 coordinates multiple cell survival pathways [2], we assessed the number of apoptotic cells after DSS treatment. In line with previous reports [21,22], Pacs2-/-intestinal epithelial cells exhibited a marked increase in the number of apoptotic cell ( Figure 3E and 3F). For further evaluation of tissue damage and inflammation, mRNA levels of the inflammatory genes tumor necrosis factor (Tnf)α, interleukin (Il)-6 and chemokine (C-X-C motif) ligand (Cxcl)1, as well as the EGFR ligands Areg, Ereg and Tgfα were analyzed by qRT-PCR. All transcripts, except Tgfα were upregulated on day 6 in the colons of both control and Pacs2 -/-DSStreated mice, with no differences in expression of any of the genes between genotypes (Supplementary Figure  1A and 1B). Likewise, expression of the fibrotic marker collagen type 1 (Col1)a1 was equally increased in Pacs2 -/and control mice after DSS treatment (Supplementary Figure 1C). These findings were supported by ELISA data, demonstrating comparable AREG and TNFα levels in supernatants from tissue explant cultures established from mice sacrificed at day 6 of the DSS-induced colitis model (Supplementary Figure 1D).

Pacs2-deficiency has no apparent effects on intestinal tumor growth in the Apc Min mouse model of colorectal cancer
Another process that relies on EGFR signaling is the development of intestinal tumors. We therefore evaluated the effects of Pacs2-deficiency in the Apc Min mouse model of colorectal cancer. The successful generation of Apc Min/+ Pacs2 -/mice was confirmed by western blot ( Figure 4A) and qRT-PCR analysis ( Figure 4B). PACS-2 was equally expressed in tumor and non-tumor tissues from the colons of Apc Min/+ Pacs2 +/+ (control) mice as shown by western blot (Figure 4A), and confirmed by qRT-PCR analysis ( Figure 4C). In contrast to our findings in the DSS-induced colitis model ( Figure 1B), Pacs2-deficiency did not affect the small intestinal expression of PACS-1 ( Figure 4B). Moreover, there were no differences in the overall colonic APC-regulated β-catenin expression levels between littermate control and Apc Min/+ Pacs2 -/mice ( Figure 4A).
Gross examination of isolated ileal tissue from control and Apc Min/+ Pacs2 -/mice revealed no obvious differences in tumor load ( Figure 5A). This was confirmed by quantification of the average total number of adenomas per mouse ( Figure 5B). We found a similar distribution of tumors in Apc Min/+ Pacs2 -/versus control mice, based on evaluation of the average number of adenomas in the proximal, middle and distal segments of the small intestine ( Figure 5C) and entire colon ( Figure 5D). Likewise, there were no changes in total tumor area per mouse ( Figure 5E) or in tumor size, quantified as average adenoma diameter ( Figure 5F).

PACS-2 loss does not affect tumor morphology, β-catenin expression or cell proliferation in the Apc Min model of colorectal cancer
H&E staining of histological sections of adenomas in the distal small intestine revealed no obvious morphological changes between control and Apc Min/+ Pacs2 -/mice ( Figure 6A). To confirm constitutive Wnt signaling in tumors upon APC mutation, we performed immunohistochemistry for β-catenin. β-catenin was localized to the perinuclear region in tumors, but cell membrane-associated in the surrounding non-tumor tissue, with no apparent differences between control and Apc Min/+ Pacs2 -/mice ( Figure 6B). Also, Ki67 immunostaining showed comparable cell proliferation in both genotypes ( Figure 6C).

DISCUSSION
We recently reported that PACS-2 regulates cellsurface availability of ADAM17, thereby fine-tuning ADAM17 activity and downstream ErbB signaling [3]. We observed a marked reduction in phosphorylated EGFR levels and proliferative index in the small intestinal epithelium and to a lesser extent in the colonic epithelium of Pacs2 -/mice. Furthermore, Pacs2 -/mice had abnormal intestinal histology with a decrease in villus/crypt ratio of the small intestine [3]. Interestingly, intestinal Loss of crypts in large areas - Histology score = E+I. The scoring system is adapted from that described by Obermeier et al. [20]. www.impactjournals.com/oncotarget  abnormalities have been reported in a number of studies of genetic alterations in the ADAM17/EGFR axis [9,[23][24][25][26][27][28]. While these phenotypes are not directly comparable to those observed in Pacs2-deficient mice, we found the similarities interesting. Apart from maintaining normal intestinal homeostasis, ADAM17 and ErbB signaling is of great importance for intestinal epithelial regeneration, and in the development of intestinal tumors [7-11, 16, 28, 29]. Here, although we observed a slight effect of Pacs2-deficiency in the context of intestinal inflammatory damage, loss of the ADAM17 fine-tuning function of PACS-2 was not sufficient to influence tumor formation driven by excessive Wnt signaling in the Apc Min model of colorectal cancer. Several studies have demonstrated that Adam17 and Egfr-deficient mice are more susceptible to colitis induced by oral administration of DSS. Typical findings include increased weight loss, higher mortality, and more severe inflammation and ulceration in these animals [8][9][10][11]. We observed a tendency towards poorer overall survival and slightly increased disease score in the Pacs2 -/mice, but none of the changes were statistically significant. Also, while loss of ADAM17 has been shown to significantly impact both inflammation and regeneration in the DSSinduced model of colitis [8], PACS-2-deficient animals only displayed indications of delayed regeneration in this model. This modest impact of Pacs2-deficiency likely results from its fine-tuning effects on ADAM17, whereby loss of PACS-2 only decreases ADAM17 cell-surface levels and intestinal EGFR activation by approximately 50% [3]. Thus, PACS-2 loss cannot be expected to have as severe effects as loss of ADAM17 or EGFR. Indeed, although we did find significantly reduced pEGFR levels in the colon of Pacs2-/-mice as compared to control mice at day 10 of the DSS-induced colitis model, no significant difference in the number of proliferating colonic epithelial cells was observed. Corroborating these marginal effects, the previously reported reduction in EGFR phosphorylation and intestinal cell proliferation in PACS-2-deficient mice was much less pronounced in the colon as compared to the small intestine [3].
We observed no apparent effects of Pacs2-deficiency in the Apc Min model of colorectal cancer when assessing tumor number, size or histology. While EGFR signaling is strongly implicated in the development of intestinal tumors [16], it appears that loss of the fine-tuning effects of PACS-2 is also not sufficient to impede tumor development or progression. It is possible that the strong proliferative signaling resulting from APC-deficiency overrides the modest anti-proliferative effects caused by the reduction in EGFR activity in Pacs2 -/animals. Additionally, it has been shown that systemic overexpression of ADAM17 in mice does not lead to enhanced shedding activity or any overt abnormalities [30], and expression of only 5% of normal ADAM17 levels is enough to overcome the embryonic lethality associated with complete ADAM17 knockout in mice [8]. Hence, the amount of ADAM17 appears not to be rate-limiting, as long as it is above a certain threshold. It is therefore likely that the fine-tuning effects of PACS-2 serve to uphold a homeostatic balance in ADAM17 levels, and that the loss of this has little functional consequence when the inflammatory or proliferative signals are very potent.
Worth mentioning, we found a small but statistically significant increase in Pacs1 expression in the colon of DSS-treated Pacs2-/-mice as compared to control mice. With a high degree of homology and frequently shared cargo of PACS proteins, this could indicate a potential functional compensation by PACS-1 upon loss of PACS-2 expression. Yet, no difference in PACS-1 mRNA levels were detected in the small intestine of control and Apc Min/+ Pacs2 -/mice. Whether PACS-1 exerts similar regulatory effects on ADAM17-mediated EGFR activation as PACS-2 is a present unknown. For future studies, it would be interesting to investigate such potentially redundant functions as well as the phenotypic impact of deleting both genes.
Moreover, PACS-2 serves numerous other functions unrelated to ADAM17 and ErbB signaling. The loss of these other functions may serve to promote inflammation or tumor growth, and in this way neutralize any antiinflammatory or anti-tumorigenic effects caused by reduced ADAM17 and ErbB signaling. Importantly, PACS-2 exerts highly complex regulatory functions on multiple cell death pathways (for a recent review see [2]). For example, PACS-2 is a key effector in TRAILmediated apoptosis, and loss of PACS-2 was shown to desensitize human colon cancer cells to TRAIL-induced apoptosis [31]. On the other hand, PACS-2 promotes cell survival following DNA damage, as exemplified by increased radiation-induced apoptosis in Pacs2 -/thymocytes [21,22]. In line with the later findings, we observed a significant increase in the number of apoptotic mucosal cells in the colon of DSS-treated Pacs2 -/mice as compared to control mice.
Although the sample size was too small to determine the relative importance, it has been reported that the human PACS2 locus is susceptible to loss of heterozygosity in colon cancer patients [32]. Also, a small selection of clinical material showed loss of PACS-2 in 4 out of 8 tumor samples, with PACS-2 expression being preserved in the surrounding non-tumor tissue [31]. While this finding needs validation in a larger patient cohort, a selective pressure for the loss of PACS-2 expression in tumor cells could indicate that the pro-apoptotic functions of PACS-2 may dominate effects on tumor formation and growth.
In conclusion, our data show that the effect of PACS-2 on ADAM17-mediated EGFR activation has negligible impact on DSS-induced colitis and the Apc Min model of cancer. These observations suggest that as a finetuner of ADAM17 and ErbB activities, Pacs2-deficiency has no notable outcome on intestinal epithelial pathologies where the induction of inflammation or cell proliferation is sufficiently strong and the amount of ADAM17 not ratelimiting.

Reagents and antibodies
DSS (Molecular weight 36,000-50,000 kDa) was obtained from MP Biomedicals (catalog no. 02160110). Hemoccult Test Kit was obtained from Beckman Coulter. Complete EDTA-free protease inhibitor cocktail was from Roche.

Animals
All animal studies were approved by the Danish Animal Inspectorate (License no. 2014-15-0201-00237). The authors confirm that all the experiments were performed in accordance with Danish guidelines and regulations for scientifically and ethically approved experimentation.
All mice were maintained in a climate-controlled room at a temperature of 22 ± 2°C and a relative humidity of 50 ± 5% under a 12-hour light/dark cycle and fed a standard diet. Pacs2 -/and Pacs2 +/+ littermate mice were co-housed and all handling of the mice, e.g. changing of cages, was done under non-sterile conditions to ensure similar microbiota in individual cages. Pacs2 -/mice on the C57BL/6 background have been previously described [31] and were from The Jackson Laboratory. Apc Min/+ mice (C57BL/6 background) were acquired through The Jackson Laboratory. Apc Min/+ Pacs2 -/mice were obtained by crossbreeding between Apc Min/+ mice and Pacs2 -/mice. The founder mice were viable and exhibited normal growth. All genotyping was performed by polymerase chain reaction (PCR) using genomic DNA prepared from mouse-tails, using the primers listed above.

DSS-induced mouse model of colitis
Littermate groups of male Pacs2 +/+ (control) and Pacs2 -/mice (8 to 16 weeks old) received 2.3% DSS (molecular weight 36,000-50,000 kDa, MP Biomedicals) in their drinking water for 5 days, followed by 5 days on regular drinking water. The optimal DSS concentration was carefully determined through several dose-response experiments starting with 2% DSS, which was the dose used in the study of hypomorphic Adam17 ex/ex mice [8]. The final experiment with 2.3% DSS was repeated three times, including a total of 21 control and 17 Pacs2 -/mice. Mice were observed and weighed on days 0, 2, and 5-10. There was no significant difference in initial body weight between the two groups. Percentage weight was calculated as (weight on a given day / weight on day 0) x 100. Stool consistency and stool blood content were monitored on days 5-10 (see Figure 1B). Mice were sacrificed upon reaching the humane endpoints defined as more than 20% weight loss, severe pain or severe rectal bleeding. All surviving mice were euthanized on day 10 and entire colons were isolated and processed for histological analysis (see below).
Weight loss, stool consistency, and stool blood content were scored as described in Table 1, adapted from the method first described by Cooper et al. [12]. Stool blood content was first determined by visual inspection of stool. If no gross blood was visible, the presence of occult blood was determined using a Hemoccult Test Kit (Beckman Coulter). The DAI was calculated as the average of the three scores [12]. In the few cases where it was not possible to acquire a stool sample, the weight loss score was substituted for the DAI. To assess tissue damage, we employed a method adapted from that previously described [20]. Histological sections of entire colons were scored blindly and independently by two researchers, one of them a trained pathologist, and subsequently compiled. Epithelial damage and extent of inflammation were scored using the criteria outlined in Table 2. An overall histology score was then calculated as the sum of the epithelial damage score and the inflammation score [20].

Analysis of intestinal tumors in the Apc Min mouse model of colorectal cancer
Eighteen-week old littermate Apc Min/+ Pacs2 +/+ (control, n = 22) and Apc Min/+ Pacs2 -/-(n = 19) mice (mixed sexes) were sacrificed, and the entire intestine was excised and divided into colon and three equal segments of small intestine: proximal, medial, and distal representing duodenum, jejunum and ileum, respectively. All regions were opened longitudinally and examined macroscopically for number, size and location of adenomas, using a stereomicroscope. Tissues were subsequently processed for histological analysis (see below).

Histological analysis
Intestinal tissues were fixed in 4% paraformaldehyde (Sigma-Aldrich) dissolved in phosphate-buffered saline (PBS), embedded in paraffin, and sectioned to a thickness of 3.5 μm. H&E and IHC staining for β-catenin and Ki67 were performed as previously described [3]. The sections were scanned using a NanoZoomer 2.0-HT Digital slide scanner C9600 (Hamamatsu) and analyzed with NDP. view 2 software. Ki67 positive proliferating cells were quantified as the percentage diaminobenzidine (DAB) positive area per total hematoxylin stained nuclear area, using ImmunoRatio, a publicly available application for quantitative image analysis [33]. Activated EGFR levels were assessed as described previously [3], using an antibody against EGFR phosphorylated on Tyr1068 and Alexa Fluor 488 conjugated secondary anti-rabbit antibody, and measuring the fluorescence intensity along a line across colonic crypts. Apoptotic intestinal epithelial cells were detected using the ApopTag ® Peroxidase In Situ Oligo Ligation (ISOL) Apoptosis Detection Kit as recommended by the manufacturer (Chemicon), and quantified using Axiovision Software (Zeiss) and manual counting of stained cells per mm 2 tissue from 3 randomly chosen mucosal areas. For quantifications, colon sections from 4-5 controls and 4-5 Pacs2 -/mice were used.

Western blotting
Intestines were opened longitudinally, and macroscopic tumors as well as non-tumor tissue was isolated. Protein extracts from non-tumor and tumor tissue were prepared using RIPA buffer (50 mM Tris pH 7.4; 150 mM NaCl; 0.1% SDS; 1% Triton X-100; 0.5% sodium deoxycholate; 1 mM EDTA, with protease inhibitor cocktail), and adjusted to equal concentration. Expression of PACS-2 and β-catenin were subsequently evaluated by western blot analysis using standard procedures, as described [3].

Ex vivo cultures of colonic explants
Tissue explants sized approximately 1 cm were harvested from the proximal colon and washed twice in cold PBS containing 100 µg/ml of streptomycin and 100 U/ml penicillin (with incubation for at least 10 min). After washing the tissue specimens, cultures were placed in a 24-well plate in 1 ml of DMEM with 100 µg/ml of streptomycin and 100 U/ml penicillin. After 24 hours, supernatants were collected, centrifuged at 1000 x g to eliminate debris, and analyzed by ELISA.

ELISA
Soluble TNFα and AREG levels were measured by ELISA in cell-free ex vivo colonic tissue culture supernatants according to the manufacturer´s instructions. Absolute values were normalized to tissue weight of colonic culture explants used in the experiment.

Statistical analysis
Unless otherwise stated, all data represent the mean ± standard error of the mean (S.E.M.). Data were analyzed by unpaired Student's t-test, analysis of variance (ANOVA) with Tukey's post-test, or log-rank (Mantel-Cox) test, as appropriate, using GraphPad Prism version 6.00 for Macintosh (GraphPad Software). In all cases, p < 0.05 was considered statistically significant.

Author contributions
S.L.D., J.S., K.J. and M.K. planned the study. S.L.D., J.S., and J.S-P. designed and performed the experiments. All authors were involved in the analysis of data. S.L.D. and J.S. wrote the manuscript, with input from M.K, and all authors reviewed the manuscript.