Collaborating clinicians must be made aware of the constraints imposed by working with RNA, and the need to preserve it so it does not ever fragment after extraction of these short ~20 nucleotide-long molecules from human stool. After consenting prospective individuals when they report to the clinic for consult, those individuals (age 18 to 90 years old) not showing polyps, or inflammatory bowel disease such as colitis or diverticulitis, will be asked by their physicians if they wish to participate in the study. If they agree, they or their guardian will be consented, each given a stool collection kit and detailed oral and written collection instructions. Each study subject will collect one 10 g stool sample, in a standardized fashion, in a large 40 cc plastic jacket prior to any bowel preparation. The study nurse will show and ask participants to obtain samples using a clean plastic spoon from both the mucinous layer, which is rich in colonocytes, and the non-mucinous parts of stool in order to have a representation of the entire colon (both right and left side colon) 335155, to be preserved in sterile urine vials overnight at room temperature in the fixative RNALater® (Invitrogen) added at 2.5 ml per 1 g of stool, followed by calling the laboratory personnel to pick up the sample by next morning. Preserved samples will be stored in our labs at -80^oC in small aliquots until needed. When ready for analysis, samples will be defrosted at room temp, filtered through a nylon mesh by laboratory personnel at GEM Tox facilities in order to remove the preservative and any debris prior to extraction of total RNA. All laboratory work is carried out and standardized under blind conditions and, in accordance with our institute s Standard Operating Procedures (SOPs) for the handling of biohazardous material.

MiRNA data are to be analyzed by RCA, and using 80% power in a Power Analysis calculation to detect differences in expression between CC stages and control group at a larger standard deviation and a 75% reduction in the difference, to optain adequate group size of control subjects and colon cancer patients from different stages of colon cancer (adenomas, TNM stage 0-I; TNM stage II; and TNM stages III & IV) selected randomly is an adequate number of samples to test. for the purpose of this study.

An adequate number of subjects who do not have colon cancer need to be chosed randomly.To avoid bias, and ensure that biomarker selection and outcome assessment does not influence each other, a prospective specimen collection retrospective blinded evaluation (PRoBE) design randomized selection 54 of control subjects and case patients from consented cohort population are chosen without knowing a priori who has what diagnosis, and will want to collect the stool specimen prior to removal of the lesion, in addition to collected consenting specimens on patients undergoing colonoscopy to form a suitable cohort to randomly selected the appropriate number of colon cancers. Then adenoma cases are matched 1 to 1 to the cancer cases for age (+/- 5 years), gender, clinic and month of diagnosis. Similarly, the normal controls from among the collected specimens are matched to the cancer and adenoma cases. If there is no match, the data are liberalized to allow +/- 2 months, in order to have collected a case-case-control group nested in the overall colonoscopy cohort that is collected. The quantitative miRNA analysis by RCA is then carried out on all coded samples at once, with the investigators blind to knowledge of the patients diagnosis, so that no analytical bias is introduced. While there may be some volunteer bias present, which may affect the studied miRNA markers, it is recommended to collecte demographic and clinical data on both groups (those who participated and those who did not) and compare for the following factors: age, gender, race/ethnicity, reason for colonoscopy, diagnoses, so that an assessment can be made at the conclusion of the study as to what degree selection may have affected study results.

A six aims and a proposed timeline for achieving these aims in a 5 year clinical research study are detailed in Table 1 below:

^a· = Refers to potential frequency and/or level of effort needed to accomplish/complete project aim.

1. Standardize sample acquisition, processing and handling; justify number of selected subjects & epidemiologically randomize selection of study population.

2. Standardize total RNA extraction from stool colonocytes by strict quality assurance (QA) criteria, and perform Rolling Circle Amplification (RCA) to study microRNAs gene expression at various stages (0- IV) of colon cancer (CC) progression.

3. Employ epigenetic methods to study genetic heterogeneity by identifying MSS & MSI phenotypes, as well as investigating promoter methylation in miRNA genes at various stages of CC Progression

4. Finalize accessing test performance characteristics of the proposed RCA diagnostic approach for colon cancer.

5. Use statistical methods for data analyses.

6. Provide and carry out alternate standardized technical methods for achieving the above aims in the unlikely event that the proposed approach, or the outlined methods fail to achieve study goals.

RNA isolation procedures (both automatic and manual), compared to those used for isolation of DNA from stool samples 5472737475767778, are standardized and made simple by us using improved commercially available reagents and kits 335179 to extract high quality total RNA from an environment as hostile as stool 80; thus, shattering the myth that it is difficult to employ RNA as a screening substrate. The trick is to stabilize total RNA shortly after obtaining fresh stool by fixing samples in a chaotropic agent and observing that RNA does not ever fragment thereafter. Fragmented RNA results in poor cDNA synthesis, and ultimately in less than optimal RCA amplification.

Total RNA can be manually isolate from colon laser capture microdissected (LCM) tissue, and from stool by using Qiagen s RNeasy Isolation Kit^® (Qiagen, Valencia, CA) containing a RLT buffer (a guanidinium-based solution) according to manufacturer s instructions, getting the advantage of manufacturer s established validation and quality control standards, thereby increasing the probability of good results. Generally, total RNA isolated from stool is suitable for amplification of miRNAs by RCA without the need to further purify mRNA because purified mRNA involves additional steps, and the increased sensitivity could be balanced for by the potential loss of material 67.

Isolated RNA in nuclease-free water can be stored at -80^oC until needed. It can then be quantified spectrophotometrically at 260 nm. Acquiring sufficient mRNA to analyze from stool or isolated colonocytes is feasible, as each cell contain ~ 20 pg total RNA, and only few nanograms are needed per RCA reaction 81.

A total of 50 pmol of forward and reverse oligonucleotides (Invitrogen), according to the sequence of the interrogated miRNA, can be annealed by incubation at 75^oC for 5 min, and then slowly cooled to room temperature (~ 30 min). The fill-in reaction to form dsDNA template can be performed in a 20 µl volume containing 10 mM Tris-HCl, 50 mM NaCl, 10 mM MgCl₂, 1 mM DTT, 0.25 mM dNTPs and 5U Klenow Fragment (3 - 5 exo) ^{New England Biolabs} at 37^oC for 1 h. Then the reaction mixture can heated at 75^oC for 20 min to inactivate the enzyme, and slowly cooled to room temperature for dscDNA annealing (see Figure 1 A, upper level).

A total of 20 µl of the above dsDNA template mixture can be added into 30 µl of in vitro transcription buffer containing 0.5 mM NTPs, 40 mM Tris-HCl, 6 mM MgCl₂, 10 mM dithiothreitol (DTT), 2 mM spermidine, 100 U ribonuclease inhibitor and 50 U T7 RNA polymerase (New England Biolabs). The reaction can be run at 37^oC for 4 h, then 1 U of RNase-free DNase I is added to digest DNA template, followed by purification of precursor miRNAs in transcription mixture with phenol/chloroform extraction. Transcript precursor miRNAs can be examined by running 2% agarose gel, and the concentration is determined spectrophotometrically at a wave length of 260 nm.

Before miRNA is reverse transcribed, the RT primer is to be chemically phosphorylated at the 5 end by heating a 0.5 nmol RT primer to 75^oC for 5 min, then chilling on ice prior to treatment with kinase. The 50 µl reaction volume containing 50 mM Tris-HCl (pH 8.0), 10 mM MgCl₂, 5 mM DTT, 2 mM DTT, 1 mM ATP, 20 U T4 polynucleotide kinase and DNA probe formation can be carried out at 37^oC for 2 h, followed by inactivation of the T4 polynucleotide kinase at 65^oC for 20 min.

RT reaction for formation of cDNA from miRNA can be carried out in a 10 µl reaction volume that contains 1µl of total RNA sample, 500 nM 5 -phosphorylated RT primers, 20 U MultiScribeO RT (Applied Biosystems, Foster City, CA) , 50 µM dNTPs and 1X reaction buffer (pH 8.3, 50 mM Tris-HCl, 75 mM KCl and 3 mM MgCl₂), followed by the addition of 2.5 U Ribonuclease H at 37^oC for 20 min for the degradation of miRNA strand in the RNA-DNA hybrid. These steps can be followed up by probe ligation and rolling circle amplification (see Figure 1 B and Figure 2).

Jonstrup et al24 (Figure 2) developed a linearly amplified simple detection system by using miRNA as a template to cyclize padlock probes as a primer for RCA, achieving a detection limit of ~ 10 pm. However, while the method is simple, linear amplifications by itself cannot satisfy the requirements for high detection sensitivity. A T7 exonuclease-assisted cyclic enzymatic (CEAM) amplification is based on nucleases in which one target leads to many cycles of target dependent nuclease cleavage of reporter probes for output signal amplification. Exonuclease III (Exo II), for example, catalyzes the stepwise removal of mononucleotides from the 3 -blunt or recessed terminus of duplex DNA. This unique property of Exo III-based CEAM allowed it to be widely used for optical (fluorescence, UV-Visible, SPR) and also for electrical amplifications for the detecting DNA, proteins and small molecules 82. However, because of the linear nature of the amplification process, it was able to only achieve detection limits in the pico molar (pM) range. Additionally, the limited intrinsic properties of the nucleases used have made earlier CEAMs not applicable to miRNA targets 83.

For extraction of genomic DNA from stool in a LRS, a manual extraction for stool samples that employs a QIAamp DNA Stool Mini Kit (Qiagen, Valencia, CA), can be used.

A fast, convenient method for label-free detection of amplified markers, which can be routinely carried out, using genomic DNA isolated from stool colonocytes by the DNA Extraction Kit (Qiagen), on 6 of the 10 Bethesda markers recommended by NCI 122. These include: three mononucleotide repeats (BAT-25, -26 and -40) and three dinucleotide repeats (D2S123, D5S346 and D17S250), which have shown a sensitivity of >96%, and a specificity of >99% for identifying tumors with MMR deficiency. The presence of MSI-H correlates with an older age of diagnosis, the presence of tumors in the proximal colon and female sex. A cut of 30% for markers has been recommended to use for classifying tumors with MSI-H 104. Label-free amplification with reagents found in the Gold Taq Kit (Applied Biosystems) can be used, as details in references 123124. Upon completion of the non-PCR amplifications in foam cups or boxes, the products of each individual sample can be processed by on-chip electrophoresis using a DNA 1000 Lab Chip in the Agilent 2100 Bioanalyzer 125126.

Modifying the genomic DNA before non-PCR amplification by sodium metabisulfite treatment 127 can be employed, followed by DNA recovery using a modified Qiagen Viral RNA Mini Kit 128, and eluted DNA samples can be stored at -80^oC until needed. Methyl Light analysis can performed by fluorescence-based assay, using β-actin as a reference normalization gene. The % of fully methylated molecule at a specific locus, called PMR, is then calculated 129130. SAS software can be used to determine the statistical significance of the results. The two major molecular CC subgroups (MSS and MSI) can be studied in selected miRNAs, at various stages of CC progression in stool samples as a complex molecular signature using factor analysis (FA) 131, combined with linear discriminate analysis (LDA) 132 to identify the cancer s molecular characteristics.

If the difference in miRNA gene expression between healthy and cancer patients and among the stages of cancer at the end of the study is as large and as informative for multiple miRNA genes, suggesting that classification procedures could be based on values exceeding a threshold, then sophisticated classification would not be needed to distinguish between the two study groups. However, if inconsistent differences on large samples are found, then it is recommended to use predictive classification methods, as detailed below.

The goal in predictive classification will be to assign cases to predefined classes based on information collected from the cases. In the simplest setting, the classes (i.e., tumors) are labeled cancerous and non-cancerous. Statistical analyses for predictive classification of the information collected (i.e., microarrays and qPCR on miRNA genes) attempt to approximate an optimal classifier. Classification can be linear, nonlinear, or nonparametric 140141. The miRNA expression data are to be analyzed first with parametric statistics such as Student t-test or analysis of variance (ANOVA) if data distribution is random, or with nonparametric Kruskall-Wallis, Mann-Whitney and Fisher exact tests, if distribution is not random 142. If needed, complicated models as multivariate analysis and logistic discrimination 143 can also be employed.

False positive discovery rates (expected portion of incorrect assignment among the expected assignments) can also be assessed by statistical methods 144145146147, as it could reflect on the effectiveness of test because of the need to do follow up tests on false positives. The number of optimal miRNA genes needed to achieve an optimum gene panel for predicting carcinogenesis in stool can also be established by statistical methods.

For the corrected index, cross-validation 148 can be used to: protect against overfitting, address the difficulties with using the data to both fit, assess the fit of the model, and determine the number of samples needed for a cancer study, where the expected proportion of genes expression common to two independently randomly selected samples is estimated to be between 20% and 50% 144. Efron and Tibshirani 149 suggested dividing the data into 10 equal parts and using one part to assess the model produced by the other nine; this is repeated for each of the 10 parts. Cross-validation provides a more realistic estimate of the misclassification rate. The area under the ROC curves, ^{in which sensitivity is plotted as a function of (1 - specificity)}, can be employed to describe the trade-off between sensitivity and specificity 150151.

Power analysis can be implemented for estimating sample size in such a study 152. Moreover, power analysis, as well as first and second order validation studies can be carried out to access the degree of separation and reproducibility of our data 153.

Principal component analysis (PCA) 154, which is a multivariate dimension reduction technique to simplify grouping of genes that show aberrant expression from those not showing expression, or a much reduced expression, can be employed. In cases where several genes by themselves appear to offer distinct & clear separation between control or cancer cases in stool samples, a predictive miRNA Index (PMI) may not be needed.

If by the end of the study, the miRNA gene panel (or a derived PMI) is better than existing screening methods, all of the data generated can be used to assess the model so over-fitting is not a concern. The level of gene expression can be displayed in a database using parallel coordinate plots 155156 produced by the lattice package in R (version 2.9.0, The R Foundation for Statistical Computing, http://cran.r-project.org), and S-plus software (Insightful Corporation, Seattle, WA). Other packages such as GESS (Gene Expression Statistical System) published by NCSS (www.ncss.com) can also be employed, with each subject having his or her medical record number as the key ID for merging various tables in the database.

If results show that individual miRNA genes offer distinct and clear separation between control and cancer, there will be not need to derive a predictive miRNA index (PMI) 157158. It may, however, be necessary to do so if data were not supportive. In this case, results of the quantitative expression of the miRNA genes can be used to determine a cutoff for a positive result. PMI results above or below the cutoff are to considered positive or negative, respectively, in all subsequent studies. Resulting data can then be then used to check the sensitivity and specificity of the index using a two by two matrix (Table 2) 159. If the sensitivity falls below 90% or the specificity falls below 95%, forthcoming data using additional miRNA genes can then be used with linear or logistic discriminant analysis to refine the index.

% Sensitivity = TP x 100

TP + FN

% Specificity = TN x 100

FP + TN

Above methods represent the most practical, least labor-intensive and economical approach to accomplish research aims. However, in few problematic samples (<5%) in control, or pre-malignant or malignant cases, it may be necessary to use other methods. However, because the error rate is so small and would occur in control and cases, adopting different extraction/analyses methods will not bias results.

In very few samples, inhibitors present in stool may make it difficult to isolate RNA using Qiagen kits that provide the advantage of manufacturer's established validation and QC standards. In such cases, RNA can be manually isolated by a modification of the classical acid guanidinium thiocyanate-phenol-chloroform (AGPC) method 160, using chaotropic guanidinium thiocyanate (GSC) that inactivates ribonucleases and most microorganisms.

Signosis, Inc., Sunnyvale, CA (www.signosisinc.com) introduced high throughput plate assay for monitoring individual miRNAs, without the need to carry out a RT reaction. In that assay one of the bridge oligos partially hybridizes with the miRNA molecule and the capture oligo, and another bridge forms a hybrid between the miRNA molecule and the detection oligo. The hybrid that is sensitive to the miRNA sequence is immobilized onto a plate and detected by a streptavidin-horse radish peroxidase conjugate and chemiluminescent substrate using a plate reader. One oligonucleotide difference prevents hybrid formation; thus miRNA isoform could be differentiated.

MiRNAs are resistant to ribonucleases present in stool, probably by inclusion in lipid or lipoprotein complexes 161 in either microvessicles (up to 1 µm), or in small membrane vesicles of endocytic origin known as exosomes (50-100 nm) 162. The mechanism of release of miRNAs from exosomes and microvesicles is unclear, although an apoptotic delivery candidate is shed from cells during apoptosis 163. Exosomes released from human and murine mast cell lines were shown to contain mRNAs and miRNAs 164. MiRNAs in microvessicles regulate cellular differentiation of blood cells and certain metabolic pathways, and modulate immune functions 165. MiRNA signatures of tumor-derived exosomes function as diagnostic markers in ovarian cancer, and tumor-derived miRNA profiles are not significantly different from exosomal miRNA profiles 166. Exosomal miRNAs was extracted from stool colonocytes by differential centrifugation, followed by filtration through 0.22 µm filters to remove cell debris, total RNA extracted by Trizol & concentration measured spectrophotometrically at λ 280 167.