Environmental enteropathy (EE) is a diffuse villous atrophy of the small bowel associated with inflammatory T-cell infiltration (1–4), of unknown etiology. EE is distinct from acute enteropathies associated with gastrointestinal infections, and has eluded attempts at a unifying precipitating cause. EE is also associated with stunting of children younger than 5 years in developing countries (5). The damage to the normal intestinal architecture has been associated with a compromise in intestinal capacity to absorb macronutrients, loss of essential enzymes leading to maldigestion, and malabsorption of a variety of nutrients, thereby exacerbating nutritional deficiencies (6–10); however, the pathophysiologic mechanisms underlying EE are not known, despite its extremely high prevalence in low-income settings in the developing world (11). EE is traditionally defined by biopsy, with subtle loss of villous height, broadening of the villi, and shortening of the crypts (3,12,5); however, intestinal biopsy is impractical in children in most settings in which this disorder occurs, so little information from biopsies are available.
An indirect test, the dual sugar absorption test, has been often used to assess gut function, and as a noninvasive way to measure enteropathy. These tests use an imbibed sugar solution that is variably absorbed, but not metabolized. These sugars are cleared by glomerular filtration (ie, there is no tubular reabsorption or secretion), and the urinary sugars are quantified, yielding ratios that presumably indicate gut integrity and function (13). The most common of these ratios is lactulose:mannitol (L:M), in which the urinary recovery of mannitol reflects small bowel surface area and that of lactulose reflects damage (14,15). Limitations of the L:M are that concurrent physiological conditions can alter L:M (intestinal transit time, delayed gastric emptying, villous maturity, renal function), and its sensitivity and specificity have not been rigorously validated. Furthermore, the L:M test is cumbersome to administer because it requires a timed urine collection, and the detection of the sugars requires sophisticated laboratory support. Additional biomarkers specific to innate and adaptive immune activation, gut damage and repair, mucosal surface area, absorptive function, and barrier regulation are needed to more comprehensively and accurately diagnose EE. Techniques that assess multiple targets present in minute quantities in readily available samples may be an opportunity to study EE in children in the developing world.
Human mRNA has been detected in stool samples (16–19), and could provide a novel window on the host gut in EE. mRNA sequences specific for small intestine inflammation are likely to be present in small amounts of feces. Moreover, the majority of fecal RNA is likely to be of microbial origin. Hence, ultrasensitive nucleic acid isolation and detection methods are needed to identify relevant host mRNAs in stool.
Digital droplet polymerase chain reaction (ddPCR) is a new high-throughput PCR platform that enables quantification of nucleic acids present in very low concentrations (20). This method partitions a conventional quantitative real-time PCR reaction (qRT-PCR) into ∼20,000 water-in-oil droplets, which permits the amplification of a single-template molecule in each droplet. Droplets are streamed single-file past a FAM/VIC 2-color fluorescence detector. The proportion of positively fluorescent droplets is used to calculate the concentration of template in the reaction.
The present study endeavored to develop reliable method of quantification of human mRNA in stool from a population of children likely to have EE using ddPCR.
Subjects and Sample Collection
Stool samples and coincident L:M data were available to the research team from 70 children ages 2 to 5 years from an ongoing longitudinal study of the gut microbiota in Malawian children (Table 1) (21). The subjects were chosen to represent 2 distinct groups, 36 children with increased L:M suggestive of EE and 34 subjects with a normal L:M. The subjects’ caretakers reported no diarrhea in the last 7 days, and none of the children was known to be infected with HIV, nor had a chronic congenital condition. Caretakers gave informed consent and the study was approved by the University of Malawi College of Medicine research ethical committee and the Washington University School of Medicine institutional review board.
Participants came from families of subsistence farmers living in mud and thatch homes in rural southern Malawi; the staple crop in this region, maize, is gathered from household gardens during a single annual harvest.
L:M testing was done using a fixed dose of dissolved sugars, 1 g mannitol and 5 g lactulose, given to children who had been fasted at least 8 hours, children were asked to void immediately before consuming the sugar solution, and this urine was discarded (11,15).
Stool was collected fresh, aliquoted with a metal spatula from a diaper into 2-mL cryovials (Fisher), and flash frozen in liquid nitrogen within 15 minutes of collection. Samples were transported to the US and subsequently stored at −80°C until analysis.
Household food insecurity and individual dietary diversity scores were measured using adapted versions of the indicator guides provided by Food and Nutrition Technical Assistance Project (22,23). Animal source food was defined as foods using meat, milk, marine products, poultry, or egg ingredients.
A 200-mg aliquot of frozen stool was transferred to a 2-mL tube containing 150 mg of 425 to 600 mm glass acid washed beads (Sigma-Aldrich, St Louis, MO) and seven to ten 2.3-mm zirconium/silica beads (Research Products International, Mt Prospect, IL). One milliliter of EasyMAG lysis buffer was added and the mixture was “bead-beated” with a FastPrep-24 tissue homogenizer (MP Biomedicals, Solon, OH) for 2 consecutive 45-second runs and then centrifuged (13,200 rpm, 10 minutes). Two hundred microliters of supernatant was used for automated RNA isolation by the NucliSENS EasyMAG lysis buffer (bioMérieux, Durham, NC) following the manufacturer's instructions with onboard lysis, “Specific A” protocol, and 110-μL elution volume (24,25).
RNA was also isolated according to the method outlined by Bennett et al (18,26) using RNeasy Mini kit (Qiagen, Hilden, Germany) from a subset of samples for comparison with NucliSENSE EasyMAG RNA isolation.
ddPCR with the QX100 Droplet Digital system (BioRad Laboratories Inc, Hercules, CA) was used to quantify expression levels of 39 different mRNA transcripts in stool RNA in 22 samples (Table 2) (19). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was selected as a measure of the total human mRNA in the samples (27). Thirty-eight additional transcripts were chosen for their potential as markers of EE, based on theoretical considerations such as their role in gut inflammation, or the representation of essential functions of the small bowel. Reactions of 20 μL were prepared using 6 μL RNA (≤20 ng/μL), 10 μL ddPCR Supermix for Probes (BioRad), 0.08 μL SuperScript III Reverse Transcriptase (200 U/μL, Invitrogen Corporation, Carlsbad, CA), 0.16 μL RNase OUT (40 U/μL, Invitrogen), 1 μL 20× TaqMan Gene Expression Assay (Applied Biosystems, Carlsbad, CA), and 2.76 (simplex) or 1.76 (duplex) μL RNase-free water. GAPDH probes were VIC-labeled, a proprietary green fluorescent dye (Life Technologies, Grand Island, NY) and run in duplex with FAM-labeled (6-carboxyflourescein) interleukin-22. All of the other probes were FAM labeled and run in simplex. All of the samples were tested in duplicate.
Six primers specific for selected mRNAs (tight junction protein 1 [TJP1], regeneration islet-derived protein 4, leucine aminopeptidase 3 (LAP3), lithostathine-1-β, β-defensin 1, and interleukin-22) were chosen for evaluation in the complete set of 70 samples on the basis of differences in the distribution as assessed by the Wilcoxon rank-sum test and visualization of the scatterplots in the initial evaluation of 22 samples. The proprietary primers were purchased from Applied Biosystems (Grand Island, NY).
Droplets were generated according to manufacturer guidelines with the QX100 Droplet generator (BioRad) before cycling in a C1000 Touch thermal cycler (BioRad) at 50°C (30:00), 95°C (10:00), 40 cycles of 94°C (0:30) followed by 60°C (1:00), 98°C (10:00). Plates were held at 4°C between amplification and droplet reading. Data from the QX100 Droplet Reader were analyzed with QuantaSoft software (Kosice, Slovakia). Fluorescent droplets were deemed positive by manually set thresholds based on results from negative control wells containing RNase-free water instead of RNA. For a given assay, all of the results shared the same threshold.
A subset of 22 samples was analyzed on 2 occasions to test intraspecimen variability. Two separate 200-mg chips of stool from the same cryovial were analyzed for TJP1:GAPDH or polymeric immunoglobulin receptor (PIGR):GAPDH ratio following the procedure outlined above.
ddPCR and Quantitative Reverse Transcriptase PCR Comparison
Twelve of the RNA samples tested with dPCR were chosen for gene expression analysis by 1-step qRT-PCR for comparison with ddPCR (27). Ten-microliter reactions were prepared using 3 μL RNA, 5 μL TaqMan Environmental Master Mix (Applied Biosystems), 0.05 μL MultiScribe Reverse Transcriptase (Applied Biosystems, Inc), 0.1 μL RNAse inhibitor (Applied Biosystems, Inc), 0.5 μL Taqman Gene Expression Mix (Applied Biosystems, Inc), and 1.35 μL RNAse-free water. The same GAPDH, LAP3, and TJP1 primers were used for qRT-PCR as for dPCR. Reactions were run in a 7500 Fast Real-time PCR system (Applied Biosystems) using a temperature cycling protocol at 48°C (30 minutes), 95°C (10 minutes), 40 cycles at 95°C (15 seconds) alternating with 60°C (1 minute).
Enzyme-linked Immunosorbent Protein Assay
Stool samples with relatively high concentrations of LAP3 and PIGR mRNA, as measured with ddPCR, were chosen for LAP and PIGR enzyme-linked immunosorbent assay (ELISA). A 50-mg aliquot of frozen stool was “beaten” in a 2-mL polypropylene tube containing 150 mg of 425 to 600 μm glass acid washed beads (Sigma). The stool was suspended in 500-μL phosphate buffer saline, vortexed at 4°C for 10 minutes, and centrifuged at 4°C and 10,000 rpm for 5 minutes. The supernatant was diluted to 2.2 mg of stool per 100 μL phosphate buffer saline to equal the concentration of stool material used in a single ddPCR reaction. Specimens were analyzed with PIGR (E91074Hu) and LAP (E90536Hu) ELISA kits (Uscn Life Sciences Inc, Wuhan, China) following manufacturer's instructions.
L:M was calculated as (lactulose)/(mannitol) (total milligrams) present in an aggregated 4-hour urine collection obtained after the oral administration of lactulose and mannitol in fixed doses. L:M <0.1 is considered to be normal on the basis of measurements made in healthy individuals in Europe and North America.
mRNA expression results for each primer were standardized against a constitutively expressed transcript, GAPDH. Descriptive statistics (mean, median, SD) for each of the mRNA transcripts measured were tabulated. Comparisons between mRNA from children with normal L:M to increased L:M were made using the 2-tailed Wilcoxon rank-sum test. The Student t test was used to determine the statistical significance of differences between means.
Children included in the present study consumed a plant-based diet with limited diversity and were likely to be stunted (Table 1).
In all of the 70 samples, >20 copies of GAPDH per 200-mg stool were detected. Of the 38 transcripts chosen for initial evaluation as potential markers for EE, 24 had copy numbers >10 in all of the 22 samples (Table 2). Of the 6 potential markers measured in all of the 70 samples, regeneration islet-derived protein 4 best differentiated children with increased L:M from children with normal L:M (Table 3).
When RNA was isolated independently from 2 different 200-mg chips from the same stool specimen and tested with ddPCR, Student paired t test showed that the 2 chips were not significantly different for both TJP1:GAPDH (n = 21, P = 0.73) and PIGR:GAPDH (n = 9, P = 0.55).
Testing of the same specimen at multiple dilutions over the range of copy numbers for 2 transcripts of interest was conducted (Fig. 1). Plotting of the transcript:GAPDH yielded a strong correlation and linear fit (r2 = 0.98 for TJP1 and r2 = 0.77 for PIGR).
Comparison of the NucliSENSE EasyMAG RNA isolation method with a more standard method using RNeasy Mini kit was done in 24 samples found to have low copy numbers. The NucliSENSE EasyMAG RNA isolation isolated detectable RNA in all of the samples, whereas the Qiagen method did so only in 11 of 24.
Digital PCR and qRT-PCR were performed on 12 random stool RNA samples using the same GAPDH, LAP3, and TJP1 gene expression assays as a direct comparison of the 2 methods. The QX100 system reports transcript quantity as copies per microliter of reaction, whereas qRT-PCR gives the number of copies per reaction. To directly compare these measurements, the ddPCR concentration was multiplied by the reaction volume of 20 μL and divided by 2. ddPCR exhibited higher sensitivity than qRT-PCR on higher copy number target GAPDH (2.2 times) and lower copy number targets LAP3, TJP1(3.3 times), with a greater relative sensitivity for the latter (Fig. 2).
Leucine aminopeptidase and PIGR were tested by commercial ELISA to compare these results with mRNA corresponding to these proteins. PIGR was detected in 2 of the 24 samples tested, whereas LAP was not reliably found.
Human mRNA that is present in extremely low copy numbers can consistently and reproducibly be isolated and detected in a set of stool samples from preschool-age children from rural Malawi. This is a promising tool with which to explore potential biomarkers of EE in the future.
The data are limited in that they come from 1 population in rural Africa. Other populations with differing dietary habits may contain fecal factors, enzymatic or microbiological, that interfere with the reliable detection of messages of interest. The absence of definitive biopsy confirmation of EE makes interpretation of the data challenging, as one does not know whether an abnormality in the L:M test has misclassified the child, or whether the temporal changes in fecal mRNA are different from L:M in healing EE, or whether the production of this (or any other) biomarker is transient. This investigation was conducted with just 70 samples, and thus the results must be considered preliminary.
This novel use of fecal mRNA, rather than fecal proteins, as a biomarker of gut health or disease may obviate long-standing techincal and performance difficulties in the field of fecal biomarkers. Detection of proteins in the stool is notoriously difficult (28). A myriad of proteases, from both the gut and its microbiota, degrades many potential protein biomarkers (29,30). It is also clear that proteins are not uniformly degraded in the colonic lumen—the factors that modulate protein degradation, the amount of water in the gut, transit time of the fecal contents, types of microbiota present, pH, and co-presence of minerals cannot easily be understood or controlled when fecal samples are collected (31,32). The 2 commercial ELISAs used in the present study did not detect the proteins LAP or PIGR in the fecal specimens. This may have been because none was present, representing loss of potential biomarker during the passage of fecal material through the colon, or limited sensitivity of the assay. Transcripts are also likely degraded to some extent in the colonic lumen; however, this technique uses an “internal standard,” that is, a constitutive message, GAPDH, by which the amounts of mRNA, even if partly degraded, are normalized. Exploration of constitutive messages other than GAPDH may be useful in further optimization of this analytical methodology. This mRNA method also requires less sample for analyses, extraction is a single process, and multiple targets can be interrogated simultaneously. This technique is also more economical than commercial ELISAs.
One challenge of working with stool samples is that stool is an amalgum of inert and biological materials, many of which could interfere with PCR amplification. A process that is too vigorous in separating RNA from other substances may damage or lose too much message, whereas testing stool without isolation of the RNA has not yet been successful as it has in samples from other media. We found that the NucliSENSE EasyMAG RNA isolation was a key component for successful analyses because at least 1 other method often yielded undetectable copy numbers of potential targets.
We anticipate that mRNA analyses in stool may be useful in identifying biomarkers for EE. The technology is powerful in that multiple targets present in nanogram amounts can be assessed simultaneously. Ultimately, biomarkers may be identified that when present are diagnostic of EE, but may require multiple specimens taken at different points in time to detect, whereas other biomarkers measure extent of damage to small bowel architecture and others gauge the inflammatory response. This technique should be investigated thoroughly with this caveat in mind, to determine its utility.
1. Menzies IS, Zuckerman MJ, Nukajam WS, et al. Geography of intestinal permeability and absorption. Gut 1999; 44:483–489.
2. Campbell DI, Murch SH, Elia M, et al. Chronic T cell-mediated enteropathy in rural west African children: relationship with nutritional status and small bowel function. Pediatr Res 2003; 54:306–311.
3. Kelly P, Menzies I, Crane R, et al. Responses of small intestinal architecture and function over time to environmental factors in a tropical population. Am J Trop Med Hyg 2004; 70:412–419.
4. Ramakrishna BS, Venkataraman S, Mukhopadhya A. Tropical malabsorption. Postgrad Med J 2006; 82:779–787.
5. Lunn PG, Northrop-Clewes CA, Downes RM. Intestinal permeability, mucosal injury, and growth faltering in Gambian infants. Lancet 1991; 338:907–910.
6. Brown KH, Khatun M, Ahmed G. Relationship of the xylose absorption status of children in Bangladesh to their absorption of macronutrients from local diets. Am J Clin Nutr 1981; 34:1540–1547.
7. Carneiro Chaves FJ, Tavarela Veloso F, Cruz I, et al. Subclinical tropical enteropathy in Angola: peroral jejunal biopsies and absorption studies in asymptomatic healthy men. Mt Sinai J Med 1981; 48:47–52.
8. Chacko A, Begum A, Mathan VI. Absorption of nutrient energy in southern Indian control subjects and patients with tropical sprue. Am J Clin Nutr 1984; 40:771–775.
9. Campbell DI, Lunn PG, Elia M. Age-related association of small intestinal mucosal enteropathy with nutritional status in rural Gambian children. Br J Nutr 2002; 88:499–505.
10. Fagundes-Neto U, Viaro T, Wehba J, et al. Tropical enteropathy in early childhood: a syndrome caused by contaminated environment. J Trop Pediatr 1984; 30:204–209.
11. Galpin L, Manary MJ, Fleming K, et al. Effect of Lactobacillus GG on intestinal integrity in Malawian children at risk of tropical enteropathy. Am J Clin Nutr 2005; 82:1040–1045.
12. Campbell DI, McPhail G, Lunn PG, et al. Intestinal inflammation measured by fecal neopterin in Gambian children with enteropathy: association with growth failure, Giardia lamblia, and intestinal permeability. J Pediatr Gastroenterol Nutr 2004; 39:153–157.
13. Meddings JB, Gibbons I. Discrimination of site-specific alterations in gastrointestinal permeability in the rat. Gastroenterology 1998; 114:83–92.
14. Shulman RJ, Eakin MN, Czyzewski DI, et al. Increased gastrointestinal permeability and gut inflammation in children with functional abdominal pain and irritable bowel syndrome. J Pediatr 2008; 153:646–650.
15. Trehan I, Shulman RJ, Ou CN, et al. double-blind, placebo-controlled trial of rifaximin, a non-absorable antibiotic, in the treatment of sub-clinical tropical enteropathy. Am J Gastroenterol 2009; 104:2326–2333.
16. Leung WK, To KF, Man EP, et al. Detection of hypermethylated DNA or cyclooxygenase-2 messenger RNA in fecal samples of patients with colorectal cancer or polyps. Am J Gastroenterol 2007; 102:1070–1076.
17. Chapkin RS, Zhao C, Ivanov I, et al. Noninvasive stool-based detection of infant gastrointestinal development using gene expression profiles from exfoliated epithelial cells. Am J Physiol Gastrointest Liver Physiol 2010; 298:G582–G589.
18. Bennett WE, González-Rivera R, Shaikh N, et al. A method for isolating and analyzing human mRNA from newborn stool. J Immunol Methods 2009; 349:56–60.
19. Rashid RA, Tabata TA, Oatley MJ, et al. Expression of putative virulence factors of Escherichia coli O157:H7 differs in bovine and human infections. Infect Immun 2006; 74:4142–4148.
20. Hindson BJ, Ness KD, Masquelier DA, et al. High-throughput Droplet Digital PCR system for absolute quantitation of DNA copy number. Anal Chem 2011; 83:8604–8610.
21. Yatsunenko T, Rey FE, Manary MJ, et al. Human gut microbiome viewed across age and geography. Nature 2012;486(7402):222–7.
22. Coates J, Swindale A, Bilinsky P. Household Food Insecurity Access Scale (HFIAS) for Measurement of Household Food Access: Indicator Guide (Vol 3). Washington, DC: Food and Nutrition Technical Assistance Project, Academy for Educational Development; 2007.
23. Swindale A, Bilinsky P. Household Dietary Diversity Score (HDDS) for Measurement of Household Food Access: Indicator Guide (Vol 2). Washington, DC: Food and Nutrition Technical Assistance Project, Academy for Educational Development; 2006.
24. Dundas N, Leos NK, Mitui M, et al. Comparison of automated nucleic acid extraction methods with manual extraction. J Mol Diagn 2008; 10:311–316.
25. Loens K, Bergs K, Ursi D, et al. Evaluation of NucliSens easyMAG for automated nucleic acid extraction from various clinical specimens. J Clin Microbiol 2007; 45:421–425.
26. Kang S, Denman SE, Morrison M, et al. An efficient RNA extraction method for estimating gut microbial diversity of polymerase chain reaction. Curr Microbiol 2009; 58:464–471.
27. Bennett WE, González-Rivera R, Puente BN, et al. Proinflammatory fecal mRNA and childhood bacterial enteric infections. Gut Microbes 2010; 1:209–212.
28. Foell D, Wittkowski H, Roth J. Monitoring disease activity by stool analyses: from occult blood to molecular markers of intestinal inflammation and damage. Gut 2009; 58:859–868.
29. Peterson CG, Sangfelt P, Wagner M, et al. Fecal levels of leukocyte markers reflect disease activity in patients with ulcerative colitis. Scand J Clin Lab Invest 2007; 67:810–820.
30. Nielson OH, Gionchetti P, Ainsworth M, et al. Rectal dialysate and fecal concentration of neutrophil gelatinase-associated lipocalin, interleukin-8 and tumor necrosis factor -alpha in ulcerative colitis. Am J Gastroenterol 1999; 94:2923–2928.
31. Vermeire S, van Assche G, Rutgeerts P. Laboratory markers in IBD: useful, magic or unnecessary toys? Gut 2006; 55:426–431.
32. Langhorst J, Elsenbruch S, Koelzer J, et al. Noninvasive markers in the assessment of intestinal inflammation in inflammatory bowel diseases: performance of lactoferrin, calprotectin and PMN-elastase, CRP, and clinical indices. Am J Gastroenterol 2008; 103:162–169.