Brachyury identifies a class of enteroendocrine cells in normal human intestinal crypts and colorectal cancer

Normal homeostasis of adult intestinal epithelium and repair following tissue damage is maintained by a balance of stem and differentiated cells, many of which are still only poorly characterised. Enteroendocrine cells of the gut are a small population of differentiated, secretory cells that are critical for integrating nutrient sensing with metabolic responses, dispersed amongst other epithelial cells. Recent evidence suggests that sub-sets of secretory enteroendocrine cells can act as reserve stem cells. Given the link between cells with stem-like properties and cancer, it is important that we identify factors that might provide a bridge between the two. Here, we identify a sub-set of chromogranin A-positive enteroendocrine cells that are positive for the developmental and cancer-associated transcription factor Brachyury in normal human small intestinal and colonic crypts. Whilst chromogranin A-positive enteroendocrine cells are also Brachyury-positive in colorectal tumours, expression of Brachyury becomes more diffuse in these samples, suggesting a more widespread function in cancer. The finding of the developmental transcription factor Brachyury in normal adult human intestinal crypts may extend the functional complexity of enteroendocrine cells and serves as a platform for assessment of the molecular processes of intestinal homeostasis that underpins our understanding of human health, cancer and aging.


Quantitative real-time PCR
cDNA was generated from the total RNA prepared from 16 normal human tissues using Quantitect Reverse Transcription kit (Qiagen, #205310). Real-time PCR reactions were carried out in triplicate in total volume of 25 μl on a CFX96 Real-Time System C1000 Thermal Cycler (BioRad) using Quantifast SYBR green RT-PCR kit (Qiagen, #204154) with 1.5 μl of diluted cDNA template (equivalent to 7.5 ng RNA). Genes of interest were amplified according to the manufacturer›s directions (initial denaturation at 95°C for 5 min, 40 cycles of 95°C for 10 s and 60°C for 30 s), followed by a melting curve analysis. QuantiTect Primer Assay (Qiagen) was used for genes listed in Table 2. Sequences of the primers designed using Primer BLAST primer designing tool (NCBI) to detect Brachyury are listed in Table 3. CFX Manager™ Version 1.0 software (Bio-Rad) with default parameters was used to assess primer efficiency and specificity and to determine the threshold cycle (Ct) values. Results were normalized to a combination of two reference genes and the relative fold change in expression was computed by the ΔΔCt method.

Brachyury-overexpression
The protein-coding region of Brachyury was cloned into the eukaryotic expression vector, pFN21A HaloTag ® CMV Flexi ® Vector (Promega, G2831) containing an N-terminal HaloTag as described in the manufacturer's instructions. SW480 cells (2 × 10 6 ) were grown in T75 flasks. Cells were transfected 24 hours later with the HaloTag-Brachyury fusion construct (FuGENE HD transfection reagent (Promega, E2311) according to manufacturer's protocols. Briefly, the ratio of FuGENE HD reagent to plasmid DNA was 3:1 and 23 μg of plasmid DNA was used per flask. Cells were harvested 48 hours post-transfection.

Source of tissue samples
Patients undergoing colonic resection for colon cancer were used to obtain pathologically normal and colon cancer FFPE tissue samples. Areas of normal bowel (terminal ileum and/or right sided colon) distal to the primary tumour were sampled from the same pathological specimens. Written consent was obtained from individual patients and ethical approval was from the local research ethics committee (North Wales Research Ethics Committee-West; reference 12/WA/0042). 22 of these normal tissue samples were used to derive counts for Figure 2F. 5 patient derived CRC samples (3 colon and 2 rectal-samples samples from RICE trial-all cancers were sporadic, one sample was grade T3N1 and all others were T3N0) were used to study co-localization of Brachyury and ChgA by IF; an example is presented in Figure 5.

Supplementary Materials
Furthermore, the CRC samples that were positive for both Brachyury and ChgA (1 colon and 1 rectal) were used in Brachyury/ChgA quantification analysis (SI below). These cancers were poorly differentiated, T3N0 adenocarcinomas.

Immunohistochemistry (IHC)
The tissue samples were fixed in formalin and embedded in paraffin and IHC analysis was performed on 4 μm tissue sections. Staining was automated on a Ventana Benchmark XT machine using a standard immunohistochemistry protocol with a heat retrieval method. Antigen retrieval and antibody dilutions are listed in Table 4. 3, 3-Diaminobenzidine (DAB) was used as a chromogenic substrate, and the slides were counterstained using haematoxylin. Negative controls were omission of the primary antibody. Images were acquired on an Axio Scan. Z1 Digital Slide Scanner and ZEN software (Zeiss).

Indirect immunofluorescence
For immunostaining of tissue, 4 µm sections were deparaffinized and rehydrated by passing them three times through xylene, two times through 100% alcohol, and two times through 70% alcohol for 10 minutes each, followed by rinsing with distilled water twice. Antigen retrieval was performed using heat-induced epitope retrieval by maintaining slides at a sub-boiling temperature (98-100˚C) in 10 mM sodium citrate buffer pH 6.0 for 10 minutes. After rinsing with distilled water, the sections were incubated in 5% FBS/ 0.3% Triton™ X-100/1× PBS for 1 hour at room temperature to block nonspecific binding and permeabilize membranes. Next, the slides were stained overnight with primary antibodies (Table 5) diluted in 1% BSA/ 0.3% Triton™ X-100/1× PBS, followed by three washes with 1x PBS for 5 minutes each. After washing, the slides were incubated with Alexa Fluor secondary antibodies (Table 6) diluted in 1% BSA/ 0.3% Triton™ X-100/1× PBS for 2 hours at room temperature in the dark, followed by three washes with 1x PBS for 5 minutes each. Coverslips were mounted with Prolong ® Gold Antifade Reagent with DAPI (Cell Signalling, #8961). Slides were analyzed using an Axio Scan. Z1 Digital Slide Scanner and ZEN software (Zeiss).

ImageJ analysis of co-staining in tissue samples
In order to quantify the percentage of CRC tissue that was co-stained for both Brachyury and ChgA a region of interest (ROI) was firstly identified from the whole tissue image (selected to avoid edge effect). This was then assessed for staining of each protein by setting a threshold using the thresholding tool in ImageJ and quantifying in terms of percentage of cells above the threshold in the ROI (according to Quantitative analysis of histological staining  and fluorescence using ImageJ http://onlinelibrary.wiley. com/doi/10.1002/ar.22641/full) Using this approach, the two CRCs that were Brachyury/CgA positive gave the following:

Supplementary
CRC 1 (poorly differentiated pT3N0 adenocarcinoma of the ascending colon): 32% Brachyury; 0.74% ChgA CRC2 (poorly differentiated pT3N0 adenocarcinoma of rectum): 10% Brachyury; 0.17% ChgA Note: Cells that were ChgA positive in the cancer samples were always Brachyury positive, so 100% of ChgA cells are also Brachyury positive but only 2.3% and 1.7% of Brachyury positive cells are also ChgA positive. were assessed, all with information both for normal (total n = 121) and colorectal cancer (total n = 406) tissues. Pearson test was used to evaluate the correlation between gene expression profiles. The statistical analysis was performed using Prism GraphPad software (version 5.0a).  (A) siRNA depletion of Brachyury and western blot using the mouse monoclonal antibody from Abcam (ab140661). U-untreated H460 cells, N-negative control siRNA treated H460 cells, T5-Brachyury siRNA (Hs_T_5) treated H460 cells, T6-Brachyury siRNA (Hs_T_6) treated H460 cells, loading was controlled by α-Tubulin. Western blot images were cropped for size. (B) Detection of bands by the mouse monoclonal antibody from Abcam (ab140661) on western blot from Brachyury-Halo tag overexpression in SW480, loading was controlled by GAPDH. www.impactjournals.com/oncotarget A total of 25 crypts stained with Ki67 (red) from one patient were used to define the transit amplifying zone (zone 2) and the differentiated zone (zone 3) of the crypts. The stem cell zone (zone 1) was defined as +5 cells from the bottom of the crypt. The average crypt length was 358.6 µm. The average length of the Ki67 positive region was 210.4 µm (59%). We defined this region (minus the stem cell zone) as the transit amplifying zone, zone 2. The average length of the Ki67 negative region was 148.4 µm (41%). We defined this as the differentiated zone, zone 3.

Pairwise comparison for differences in the frequency of Brachyury positive cells in zone 1, 2 and 3:
Group 1 Group 2 P value Zone 1 Zone 2 < 0.0001 Zone 1 Zone 3 < 0.0001 Zone 3 Zone 2 < 0.0001 Supplementary Figure S3: Use of the proliferative marker Ki67 to define the zones of the normal colon crypts.