Modification of Diet to Reduce the Stemness and Tumorigenicity of Murine and Human Intestinal Cells

Scope Black raspberries (BRBs) have colorectal cancer (CRC) chemo‐preventative effects. As CRC originates from an intestinal stem cell (ISC) this study has investigated the impact of BRBs on normal and mutant ISCs. Methods and results Mice with an inducible Apcfl mutation in either the ISC (Lgr5CreERT2 ) or intestinal crypt (AhCre/VillinCreERT2 ) are fed a control or 10% BRB‐supplemented diet. This study uses immunohistochemistry, gene expression analysis, and organoid culture to evaluate the effect of BRBs on intestinal homeostasis. RNAscope is performed for ISC markers on CRC adjacent normal colonic tissue pre and post BRB intervention from patients. 10% BRB diet has no overt effect on murine intestinal homeostasis, despite a reduced stem cell number. Following Apc ISC deletion, BRB diet extends lifespan and reduces tumor area. In the AhCre model, BRB diet attenuates the “crypt‐progenitor” phenotype and reduces ISC marker gene expression. In ex vivo culture BRBs reduce the self‐renewal capacity of murine and human Apc deficient organoids. Finally, the study observes a reduction in ISC marker gene expression in adjacent normal crypts following introduction of BRBs to the human bowel. Conclusion BRBs play a role in CRC chemoprevention by protectively regulating the ISC compartment and further supports the use of BRBs in CRC prevention.


Introduction
CRCis linked to dietary choices and is the 2nd leading cause of malignancyrelated deaths in the Western world. [1] With ≈50% of cases thought to be preventable by lifestyle choices [2] there is a need to understand how diet impacts upon the normal intestine. Ultimately the interaction of an individual's diet with their intestine impacts on theirISCs which maintain the epithelial barrier and are considered the "cell of origin" of CRC. [3] Current research examining a high-fat diet (HFD) and CRC has demonstrated it increases ISC numbers and the risk of oncogenic transformation; [4] setting a precedent for the ISC impact of diet. Potentially dietary components associated with reduced CRC risk may be linked to a reduction in ISC activity in the healthy intestine. Previous studies have demonstrated that administration of BRBs inhibited tumorigenesis in: Apc1638 +/− mice, inflammation-driven CRC in Muc2 −/− mice, [5] CRC patients, [6] and lead to polyp regression in familial adenomatous polyposis (FAP) patients. [7] FAP patients inherit a mutated copy of the APC gene, APC loss in an ISC activates the WNT pathway, and is the earliest known event in CRC; a key feature of inherited and sporadic (≈90%) CRCs. In this study we used pre-clinical ex vivo and in vivo models and material from a CRC BRB clinical intervention trial to establish whether BRBs impact on CRC via modulation of the ISC pool.

Dietary BRBs Extend Survival of Lgr5CreER T2 Apc fl/fl Mice
To establish whether BRBs could impact on ISCs, cohorts of Lgr5CreER T2 Apc +/+ and Lgr5CreER T2 Apc fl/fl mice were randomly assigned to either control or 10% BRB diet from 2 weeks prior to Cre induced ISC Apc deletion ( Figure S1A, Supporting Information). Analysis of weight change demonstrated no effect of BRB administration in either cohort ( Figure S1B,C). At 190 days post induction (d.p.i.) there was no significant difference in the survival of the Lgr5CreER T2 Apc +/+ cohorts ( Figure 1A). In contrast, the BRB diet significantly increased survival of Lgr5CreER T2 Apc fl/fl mice ( Figure 1A). With the area of nuclear -catenin positive lesions (a surrogate marker for Apc loss) at 20 d.p.i (N = 4) significantly reduced in the BRB cohort (Figure 1B,C). The reduction in lesions at 20 d.p.i in BRB treated mice was not the result of reduced proliferation or increased cell death as the number of BrdU+, Ki67+ (proliferation), and CC3+ (Cleaved Caspase-3, cell death) cells were unaltered between the two cohorts ( Figure 1D-F). However, in endpoint tumors, there was no change in proliferation but a significant reduction in CC3+ cell death ( Figure 1D-F). To assess the impact of BRBs on the ISC population we used the AhCreApc fl/fl [8] acute model of Apc loss to amplify alterations to the ISCs and Wnt signaling pathway. Following induction, AhCre driven Apc loss acutely activates the canonical Wnt pathway in all cells of the intestinal crypt (except Paneth cells). Mice were placed on diets 2-weeks prior to Apc loss and harvested 5 d.p.i. In Apc +/+ mice gene expression analysis of a range of putative ISC markers indicated BRBs significantly reduced expression of the key ISC marker Olfm4 and a non-significant >4 fold reduction in Lgr5 ( Figure 1G,H). A similar pattern was observed in the AhCreApc fl/fl BRB treated mice with significant reduction in Olfm4, a decrease in Lgr5 expression and significant upregulation of Ascl2 (Figure S1D,E). The absence of significant upregulation of the Wnt target Lgr5, in this model of acute Wnt activation is consistent with an ISC role for BRBs and further supported by the increase in the Wnt target Ascl2. Ascl2 plays a role in the dedifferentiation of neighboring non-ISC epithelial cells to an ISC state; a mechanism for replenishing a depleted Lgr5 stem cell pool. [9] As a HFD promotes tumorigenesis in mice by increasing ISC number [4] we next determined whether BRBs reduced the number of normal ISCs. Using the Lgr5CreER T2 Apc +/+ mouse we demonstrated that a 2-week BRB diet significantly reduces the ratio of ISCs:non-ISCs in the normal crypt ( Figure 1I,J). Suggesting that in this model BRBs may prevent intestinal tumorigenesis by decreasing the number of Lgr5+ ISCs prior to and following Apc loss.

BRB Exposure Modifies the Apc Deficient "crypt-progenitor phenotype"
As BRBs impacted the ISC population we assessed the effect of BRBs on Apc −/− crypt cell dynamics with the AhCreApc fl/fl model. Despite a reduction in the number of ISCs, BRB exposure had no significant effect on crypt cell number, mitosis, proliferation or cell death in WT mice, except for an increase in CC3+ cells in VillinCreER T2 Apc +/+ crypts (Figure 2A-C). To determine whether the BRB diet specifically targeted the ISCs for cell death we quantified the position of apoptotic CC3+ cells in the VillinCreERT2Apc +/+ mice. The position of CC3+ cells were not different between control and BRB diet, and the majority of cell death occurred above cell position 5, suggesting that cell death mostly occurred outside the stem cell population ( Figure 2C). The "crypt-progenitor phenotype" is a term used to characterize the immediate consequences of Apc loss within the intestinal crypt. Following the loss of Apc, a rapid localization of -catenin to the nucleus results in an Wnt-activated, elongated, hyper-proliferative crypt due to an increase in ISC and Paneth cells (and mislocalization of these cell types) concomitantly with increased cell death, failed differentiation, and aberrant migration. [10] In contrast to the Apc +/+ setting, BRB diet significantly modified the Apc −/− "crypt-progenitor phenotype" (Figure 2A,B,D-H). [10] With a BRB diet further elevating the increase in total number of cells/crypts, proliferating cells and apoptotic bodies and restoring migration along the crypt-villus axis (Figure 2A,B,D-H). In the Apc fl/fl crypts, whereby loss of Apc has been shown to increase the number and size of the stem cell population [10] the position of cell death was slightly altered in BRB fed mice, such that apoptosis was shifted by approximately three cell positions lower (towards the crypt base) than that in the control fed mice, however, cell death in both dietary settings occurred equally within the first 15 cell positions ( Figure 2D), suggesting that BRBs do not selectively target crypt base columnar stem cells for cell death. Due to the increase in stem-like cells in Apc fl/fl crypts, it is plausible that BRBs push these stem-like cells to apoptosis-mediated cell death and this could aid in the attenuation of the "crypt-progenitor phenotype." The restoration of cell migration indicated that BRBs were influencing the signaling pathways that direct differentiation, both of which are inhibited in the AhCreApc fl/fl model. In WT mice, BRBs significantly increased the nutrient sensing enteroendocrine cells with no effect on goblet or Paneth cell numbers (Figure 3A-C). With BRB diet attenuating the Apc fl/fl phenotype, [10] significantly increasing enteroendocrine ( Figure 3A,E) and goblet cells ( Figure 3B,F), while Paneth cells decreased in number ( Figure 3C,G) and localize back towards the base of the crypt ( Figure 3D). This increase in the number of apoptotic CC3, enteroendocrine, and goblet cells was confirmed using the VillinCreER T2 Apc fl/fl model, which deletes Apc in the entire crypt-villus compartment, including the Wnt3A producing Paneth cells ( Figures 2B and 3A,B). As BRBs increased apoptosis in AhCre, LgrCreER T2 , and VillinCreER T2 Apc fl/fl mice we next sought to establish whether the increase in survival is due to increased Apc −/− cell death rather than a reduction in ISCs prior to Apc loss. To establish this, we used the following assumptions: 1) each Apc −/− cell has a production rate of Pr (BrdU+ cells/total cells) and an apoptotic rate Ap (CC3+ cells/total cells), and 2) these rates are the same over all cells and constant over   thus Pr-Ap = 0.36. Hence, the net growth rate of the Apc −/− population has increased from 0.21 to 0.36 on a BRB diet; in contrast to the observed reduction in area of -catenin positive lesions ( Figure 1B,C). Together this supports the rationale that in Lgr5CreER T2 Apc fl/fl mice a BRB diet reduces the number of ISCs required to maintain homeostasis, thus limiting the number of Apc −/− ISCs following induction which alongside BRB suppression of the Apc fl/fl "crypt-progenitor phenotype" manifests as reduced tumorigenesis and increased lifespan.

BRB-Derived Anthocyanins Reduce Self-Renewal Efficiency of Apc Deficient Cells
To understand the relationship between BRB diet and reduced ISC gene expression in Apc fl/fl cells in vivo we performed a stem cell functionality assay using 3D ex vivo organoid culture [11] (Figure S1F). As expected, a 2-week exposure to BRBs in the diet prior to Apc deletion, did not impair the ability of Apc fl/fl crypts to form organoids or impede their growth compared to control (Figure 4A,B), indicating that despite a reduction in ISCs ( Figure 1I,J) each crypt contains a functional ISC. Thus, we next set out to evaluate whether exposure to BRBs ex vivo influenced the viability and self-renewal capacity of Apc fl/fl organoids. We treated Apc fl/fl organoids with increasing concentrations of BRB-derived ACs (spanning novel concentrations from 15.6 to 16 000 μg mL −1 ) that have been previously used on several human CRC cell lines. [12] Cell viability demonstrated that Apc fl/fl organoids are sensitive to BRB-derived ACs in a dose-dependent fashion ( Figure 4C; IC 50 = 1.3 mg mL −1 ). One-week exposure of Apc fl/fl crypts to sub lethal low toxicity AC concentrations between 0 and 500 μg mL −1 indicated no differences in organoid forming potential but at 500 μg mL −1 organoids were significantly smaller than control ( Figure 4D,E,H); reflecting either a reduction in ISC numbers or reduced proliferative capacity. To examine this, AC treated organoids were split to single cells and reseeded for organoid culture. Following reseeding there was a reduction in the percentage of organoid forming cells in a dose dependent fashion, which was significant at 500 μg mL −1 (Figure 4F,H). There was no significant alteration to organoid growth following passage, but this could be attributed to the number of countable organoids treated with 500 μg mL −1 AC ( Figure 4G). Together this data demonstrates that BRB-derived ACs at sub-lethal levels reduce ISC number and activity.

BRB Exposure Reduces ISC Gene Expression in the Human Bowel and the Tumorigenicity of CRC Organoid Cells
As a BRB diet has previously been shown to be effective in FAP [7] and CRC patients [6] we next investigated whether BRB treatment affects human ISCs. Using the ISO48, ISO50, and Caco2 cells grown in 3D we demonstrated that, akin to the mouse, they are sensitive to increasing AC concentrations in a dose-dependent manner (Figure 5A-C; ISO50 IC 50 = 1.74 mg mL −1 , ISO48 IC 50 = 3.76 mg mL −1 , and Caco2 IC 50 = 0.6 mg mL −1 ). With the efficiency of ISO50 single cells to self-renew significantly impaired following exposure to a sub-lethal 500 μg mL −1 of ACs ( Figure 5D), irrespective of passage number ( Figure 5E,F). This reduction in ISO50 self-renewal is supported by a significant reduction in gene expression of the ISC markers LGR5, OLFM4, and ASCL2 in the organoids after a weeklong exposure to the AC extract ( Figure 5G). To corroborate our data in humans we investigated whether intervention with oral BRBs, in CRC patients, [6] altered ISC marker expression in normal CRC adjacent colonic tissues. Using RNAscope we stained for ISC markers OLFM4 and LGR5 before and after BRB intervention ( Figure 5H-J). For OLFM4, 3/4 patients demonstrated a decrease in OLFM4 staining intensity, significant in patient 19, and no alteration in patient 10 ( Figure 5H). Due to the small tissue samples and number of countable crypts, analysis of OLFM4 staining in crypts across all patients (N = 22 crypts/6 patients pre-intervention, N = 36 crypts/8 patients post-intervention) indicated a significant reduction in OLFM4 expression post BRB intervention ( Figure 5H,I). Changes in LGR5 expression were inconsistent; it was unaltered in 3/8 patients, significantly increased in 3/8 patients and significantly decreased in 2/8 patients ( Figure 5I), potentially due to the proximity of this tissue to the leading edge of a CRC affecting this Wnt target gene. Combined analysis indicated there was no meaningful change in LGR5 expression across the cohorts (N = 45 crypts/10 patients pre-intervention, N = 117 crypts/10 patients post-intervention) ( Figure 5J). As the original study was a biomarker assay, clinical endpoints were not collected. However, the stem cell marker changes were observed in patients who were exposed to BRB intervention for longer than 4 weeks in which significant improvement in prognostic biomarkers was most apparent ( Figure 5K). The OLFM4 data are consistent with our preclinical mouse data and despite low numbers indicate that BRBs impact on stem cell characteristics within the human bowel.

Conclusions
The nutritional polyphenols found in BRBs have been shown to have pleiotropic effects against CRC initiation and progression. [5][6][7]12] While there is strong evidence that BRBs reduce FAP polyp [7] and tumor burdens, [5,6] alter the tumor microenvironment, [13,14] and reduces cancer-related inflammation, [5] there are limited studies on their effect on normal and malignant ISC populations. Given the known importance of an ISC as the cell of origin of CRC, prevention of CRC, in part, must relate to the number of ISCs required to maintain homeostasis. As ISCs play a role in how tissues adapt to alterations in dietary stimuli, there is increasing evidence that diet and obesity impacts on the ISC population and CRC risk. Thus, our data indicates that dietary exposure to BRBs prevents CRC by reducing the number of normal ISCs required for intestinal homeostasis and suppressing Apc deficient ISCs. It is important to note the impact on mutated cells, as while it is entirely plausible that decreasing ISCs can lead to decreased CRC risk the picture is not clear. ISCs follow a neutral drift model, in which each ISC has an equal probability of replacing its neighboring ISC or being replaced. Apc mutated ISCs follow a biased drift model, as it has an increase in clonal fitness which favors its retention. Reducing normal ISCs has the effect of increasing the fitness of a mutated ISC, thereby increasing the chance of it becoming fixed and forming a CRC. [15] Therefore, for a reduction in CRC risk to be elicited by reducing normal ISC numbers, there must be an equivalent reduction in the fitness of any mutated ISC to prevent its increased likelihood of fixation, which we report here. It is of note then that the reduction in ISCs and ISC gene expression in BRB fed AhCreApc fl/fl mice is associated with increased Ascl2 expression ( Figure S1D, Supporting Information). Ascl2 upregulation is associated with loss of Lgr5+-ISCs [9] and may represent emergence of Ascl2 + /Lgr5 − non-ISC cells that migrate into the ISC compartment to dedifferentiate and replace the lost or damaged Lgr5+-ISCs. [9] As we demonstrated that the expansion of the ISC [10] compartment, due to Apc loss, is attenuated by a BRB diet potentially the increased Ascl2 expression may represent the generation of a stream of cells attempting to replace the functional stem cells which ultimately fails as they are continuingly being lost due to a diminished competitive advantage over normal neighbors. However, it remains to be seen whether the increase in Ascl2 expression in the AhCreApc fl/fl would lead to rapid Apc −/− ISC expansion if BRBs were removed from the diet. In the healthy AhCreApc +/+ model it is of note that Ascl2 is not upregulated upon BRB exposure suggesting that the reduced numbers of ISCs is adequate to maintain cellular and tissue homeostasis and do not require replenishment from outside the ISC compartment.
The links between obesity and increased cancer risk are becoming well established. [2] Importantly, we report that BRBs have no adverse effects on murine body weight over time consistent with other findings. [16] Obesity is a major cause of chronic inflammation, a risk factor for metabolic syndromes and diabetes. There is some evidence in metabolic syndrome [17] and pre-diabetic patients [18] that BRB interventions have positive effects on blood lipid levels and inflammation. Therefore, future studies should investigate the effects of BRBs on blood sugar levels, serum lipid profiles, and inflammatory markers in cancer. Furthermore, tumor-induced weight loss is a common feature of cancer. We report weight loss as mice become symptomatic of disease, however mice treated with BRB diet tend to have on average a higher body weight than control suggesting that BRBs may maintain healthier weight even when symptomatic of disease which may contribute to our observed improved survival, however this needs further investigation.
In this study, we report that long-term feeding of a BRB diet has no detrimental impact on overall health despite reductions in ISC number. This is in accordance with reports indicating that intestinal homeostasis is not perturbed following the complete ablation of Lgr5-positive stem cells in mice, [19,20] while the significant increase in enteroendocrine cells in mice exposed to BRB is likely due to their roles as chemo-sensors in the gut. [21,22] A previous study demonstrated that mice fed HFD had significantly lower numbers of enteroendocrine cells in the small intestine, [4] thus it is possible that this cell population may increase in response to a change in dietary stimuli, in this instance, BRB diet. In addition, in the 2011 phase 1 clinical trial studying the effects of BRB slurry in CRC patients, it was reported that patients showed differential response to the BRB intervention, which was in part related to the length of time that patients were on the BRB treatment. [6] This may account for our observed patient variation in colonic LGR5 expression. The data also suggests that prolonged exposure (>4 weeks) to BRBs impacts the ISC population, which could correlate with the positive changes in epigenetic biomarkers previously reported. [6] Additionally, it is important to note that the ISCs changes are observed in the normal adjacent tissue and thus may reflect LGR5 driven repair, therefore it would be important to distinguish whether upregulation of LGR5 expression reflects an increase in individual ISCs or an increase in Wnt activity in the in situ ISCs. Finally, questions remain over how the ability of BRBs to reduce ISCs competes with the ability of a high-fat diet to increase them, but overall, this data provides evidence that a diet high in anthocyanin-containing fruit may augment current CRC therapies and support campaigns to adopt a healthier lifestyle.

Experimental Section
Below was a summary of experiments, detailed information was available in supplementary material.
Animal Experiments: Work was approved by a UK Home Office Project licence (30/3279) and reported in accordance with institutional and NC3R(UK) ARRIVE guidelines. Mixed sex outbred C57Bl6/J mice (11-19 weeks) were used with the following transgenes: Apc fl , [23] AhCre, [8] VillinCreER T2 , [24] and Lgr5-EGFP-IRES-CreER T2 (Lgr5CreER T2 ). [25] Mice were randomly assigned to their respective diets from 2-weeks prior to Cre activation until sacrifice at either a specific time point or a humane endpoint when symptomatic of disease ( Figure S1A, Supporting Information).
Cell Analysis: For immunohistochemistry (IHC), tissue was fixed in 10% neutral buffered formalin (Sigma, UK) and processed by conventional means. The following antibodies were used to stain for: Apc deficient cells anti--catenin (Transduction Lab #610154); apoptosis anti-cleaved caspase 3 (CC3; CST #9661); for proliferation and migration anti-BrdU (BD Figure 5. BRBs alter human stem cell marker gene expression and suppress human CRC cells in ex vivo culture. AC viability curves of ex vivo human CRC organoids A) ISO50 (IC 50 = 1.74 mg mL −1 ); B) ISO48 (IC 50 = 3.7 mg mL −1 ) and C) Caco2 cells (IC 50 = 0.6 mg mL −1 ), ≥4 technical replicates each. 500 μg mL −1 AC exposure reduces the self-renewal capacity of human ISO50 organoid cells (D; N = 6 technical replicates) (E; N = 3 different passages, note one biological replicate is the average of the data in panel D) (F; representative images of human ISO50 organoids exposed to 0 and 500 μg mL −1 BRB-derived ACs for 1-week after passage to single cell, 1 mm) and suppresses expression of ISC marker genes (G; N = 3-4 different passages). CRC patients administered oral BRBs (60 g day −1 ; 1-9 weeks) have a reduction in OLFM4 expression (H; combined N = 8 patients/58 crypts and I; OLFM4 brown, 50 μm) and differential LGR5 expression (J; combined N = 10 patients/162 crypts) in CRC normal adjacent crypts. K) Table summarizing the length of time patients were on BRB intervention, showing that prolonged treatment results in changes to the ISC marker expression.
Biosciences #347580); proliferation anti-Ki67 (Abcam #ab16667); Paneth cells anti-lysozyme (ThermoScientific #RB372-A1), and ISCs anti-GFP (CST #2956S). Protocols were available upon request. Enteroendocrine and goblet cells were stained with grimelius [26] and alcian blue, [27] respectively. mRNA was visualized via in situ hybridization [28] or RNAScope. Cellular analysis was performed within small intestinal tumors or on a total of 25 crypts from the first 5 cm of small intestine. Tumor burden was determined as a ratio of the area of nuclear -catenin positivity versus the total crypt area in longitudinal sections.
Expression Analysis: RNA was isolated from 0.5 cm of frozen mouse small intestinal tissue (≈5 cm distal to stomach) or human organoids and transcribed using Superscript III (Invitrogen, UK). Relative gene expression analysis was carried out using SYBR green fast master mix (Applied Biosystems), primer sequences available on request, or TaqMan Universal PCR Master Mix with Assay on Demand probes (see Supplemental Information).
In Vitro/Ex Vivo Analysis: For stemness and self-renewal assays, the study used mouse organoids generated from intestinal crypts, [11] human CRC organoid lines ISO48 and ISO50 (Cellesce, UK), and the Caco2 cell line. For in vivo stemness assays ( Figure S1F, Supporting Information), VillinCreER T2 Apc fl/fl mice were fed for 2-weeks prior to induction. Mice were harvested 3 d.p.i and 200 crypts per well were plated in Matrigel in organoid culture medium. Medium was replaced every 2 days and organoid number determined at day 7 (Gelcount, Oxford Optronix, UK). For self-renewal efficiency, organoids (mouse N = 3-4 lines; human N = 3 different passages) were cultured with AC medium for 7 days before dissociation into single cells using TrypLE (ThermoFisher Scientific, UK). Plates were seeded with 8000 cells per well in the presence of AC [12] and ROCK inhibitor (Stem Cell Technologies, UK; first 3 days only, 10 μM final concentration). The number of organoids formed at day 7 was used to determine self-renewal efficiency ( Figure S1F, Supporting Information). For cell viability analysis, 8000 single cells were plated in normal medium and AC extract added at day 4. On day 7 post-seeding, viability was assessed using CellTiter-Glo (Promega, UK) and a CLARIOstar plate reader (BMG Labteck, UK).
BRB Clinical Trial: The clinical trial was approved by the Institutional Review Boards of the Ohio State University Comprehensive Cancer Center and the University of Texas, San Antonio. All patients accrued to the trial had a diagnosis of CRC. [6] Biopsies of CRC adjacent normal tissue (<2 mm) were obtained 24 h prior to administration of BRB. 20 g freezedried berry powder was mixed with 100 mL of water and consumed orally three times a day (60 g day −1 total (equivalent to ≈7% rodent BRB diet) 6 h -1 apart) for 1-9 weeks. Patients remained on BRBs until 12-36 h prior to surgery to remove tumor. At surgery, an additional three adjacent normal tissue biopsies were taken from each patient. All tissue specimens were fixed in formalin and assessed by a medical pathologist.
Statistics: Genotype and/or diet controls were used throughout this study and no bias was applied during husbandry, tissue sampling, or outcome analysis. On graphs if not indicated otherwise, p values are: * p < 0.05; * * p < 0.01; * * * p <0.001, with data points indicated on graph with mean ± standard deviation (SD).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.