Digestion of so-called resistant starch sources in the human small intestine

¹ From the Department of Pediatrics, Laboratory of Nutrition and Metabolism, University Hospital and University of Groningen, Groningen, Netherlands.

² Supported by grant no. STW-NWO GGN.4487 from the Dutch Foundation of Technical Sciences.

³ Address reprint requests to RJ Vonk, Laboratory of Nutrition and Metabolism, Laboratory Center CMC V, Y2147, University Hospital Groningen, PO Box 30.001, 9700 RB Groningen, Netherlands. E-mail: r.j.vonk{at}med.rug.nl.

ABSTRACT  
Background: Resistant starch sources, which are only partially digested in the small intestine, can be used to increase colonic availability of short-chain fatty acids.

Objective: To study the characteristics of the fermentation of resistant starch, the digestion of resistant starch in the small intestine has to be quantified. We compared the metabolic fates of highly digestible cornstarch (DCS), Hylon VII (type 2 resistant starch), and Novelose 330 (type 3 resistant starch), which are of corn origin and, therefore, naturally enriched in ¹³C.

Design: After administration of 40 g starch or glucose to 7 healthy volunteers, glucose and exogenous glucose concentrations in serum and ¹³CO₂ excretion in breath were analyzed for 6 h. ¹³C abundance in carbon dioxide was analyzed by isotope ratio mass spectrometry (IRMS) and ¹³C abundance in glucose by gas chromatography–combustion IRMS.

Results: By comparing the area under the curve (2 h) of exogenous glucose concentration in serum (¹³C glycemic index) after intake of starch or glucose, ¹³C glycemic indexes for DCS, Hylon VII, and Novelose 330 were calculated to be 82 ± 23%, 44 ± 16%, and 43 ± 15%, respectively. Comparison of 6-h cumulative percentage dose recovery in breath showed that 119 ± 28% of DCS, 55 ± 23% of Hylon VII, and 50 ± 26% of Novelose 330 was digested in the small intestine.

Conclusion: The exogenous glucose response in serum and the ¹³CO₂ excretion in breath can be used to estimate small intestinal digestion of resistant starch, which amounts to 50%.

Key Words: Resistant starch • glycemic index • glucose • digestion • stable isotopes • healthy subjects • adults

INTRODUCTION  
To increase the supply of short-chain fatty acids in the colon, resistant starch can be consumed (1, 2). For studying the overall biological effects of fermentation of resistant starch, a quantitative approach is desirable. However, quantitation is difficult because of the partial digestion of the resistant starch source in the small intestine. Several techniques have been used in humans to quantify the amount of starch that resists pancreatic amylase, including direct intubation of the ileum, ileostomy sampling, measurement of the breath-hydrogen response due to colonic fermentation, and in vitro digestion methods (3–11). All of these techniques have intrinsic shortcomings. Therefore, we developed an in vivo approach under physiologic conditions. For this purpose, we used 3 types of starch as model compounds: highly digestible cornstarch (DCS, Custard; AJP, Koog aan de Zaan, Netherlands), Hylon VII (Unilever Research Laboratory, Vlaardingen, Netherlands), and Novelose 330 (Unilever Research Laboratory). DCS and resistant starch vary greatly in the ratio of amylose to amylopectin (26% amylose:74% amylopectin and 62% amylose:38% amylopectin, respectively). Hylon VII and Novelose 330 differ in amylose structure: Hylon VII is a type 2 resistant starch and Novelose 330 is a type 3 resistant starch (2). All 3 types of starch are of corn origin and, therefore, naturally enriched in ¹³C (12). After intestinal absorption, exogenous glucose is ultimately oxidized to ¹³CO₂. This implies that besides the exogenous glucose response in serum, the ¹³CO₂ excretion in breath can be measured after consumption of ¹³C-enriched substrates (13–16). With advanced mass spectrometry technology, we monitored the metabolism of cornstarch–derived glucose in blood and ¹³CO₂ excretion in breath and calculated the small intestinal digestion of each of these starches.

SUBJECTS AND METHODS  
Experimental protocol
Three groups of 7 young, healthy volunteers (aged 20–26 y) consumed glucose and 1 of the 3 starches. Glucose was used as an intraindividual standard. Five volunteers participated in >2 tests. The substrates were studied in a parallel group design. For 6 h, the glucose concentration and the ¹³C-¹²C ratio in serum and the ¹³C-¹²C ratio in breath carbon dioxide were analyzed. From the serum glucose response, the glycemic index was calculated according to the method of Jenkins et al (17). From the serum exogenous glucose response, the ¹³C glycemic index was calculated (see Calculations). A digestion index was also calculated from the ¹³CO₂ excretion data (¹³CO₂ glycemic index). The protocol was approved by the Medical Ethical Committee of the Faculty of Medical Sciences of the University of Groningen.

The subjects were asked to not consume ¹³C-rich foods such as corn products, cane sugar, and pineapple for 2 d before the test. Furthermore, the subjects fasted from 2200 the evening preceding the test until the start of the experiment at 0800 the next day. During the test, they refrained from consuming food and energy-containing drinks and remained seated during the 6-h test period. The experiments were performed with intervals of 48 h between the tests to wash out residual ¹³C from the previous test. Forty grams of each substrate was dissolved in 200 mL low-fat milk 1 h before the start of the test: DCS, ¹³C enrichment relative to Pee Dee belemnite (¹³C_PDB) = -12.19; Hylon VII, ¹³C_PDB = -12.33; or Novelose 330, ¹³C_PDB = -11.44. Glucose derived from corn had a ¹³C_PDB value of -14.93. Flavorings and sweeteners (without energy) were added to make the test meal palatable. Finally, gelatin was added to the mixture to stiffen it and to get equal consistencies for raw and cooked starch test meals (not described in this article).

Blood sampling and breath collection
Two-milliliter blood samples were collected before, at 15-min intervals during the first hour, every 30 min during the second and third hour after the meal, and every hour until the end of the experimental period (6 h). For this purpose, an intravenous catheter was placed in the cubital vein of one of each subject's arms. Blood was collected in evacuated containers containing 5 mg sodium fluoride and 4 mg potassium oxalate (Becton Dickinson, Meylan Cedex, France). Blood was centrifuged at 900 x g for 10 min and plasma was stored at -20°C until analyzed. Breath samples were collected in triplicate in 10-mL gas collection tubes (Exetainers; Labco, High Wycombe, United Kingdom) at 15-min intervals during the first 2 h and every 30 min thereafter until the end of the experiment. Breath air was blown into the gas collection tubes with a straw.

Analytic procedures
Breath 13CO2 analysis
The procedure to analyze ¹³C enrichment in breath in our laboratory was described previously (18). During the 6-h period, the total carbon dioxide exhalation rate was measured 3 times at 1.5 h intervals with indirect calorimetry (Oxycon; Dräger, Breda, Netherlands).

Plasma exogenous glucose determination
After serum samples were thawed, one aliquot was analyzed for glucose concentration by applying routine techniques using an ECA-18 glucose analyzer (Medingen, Dresden, Germany) and a second aliquot was prepared for analysis of the ¹³C-¹²C ratio. The method for calculating exogenous glucose was first described by Normand et al (13); we used the same method with some modifications. To denature plasma proteins, 100 µL plasma was added to 1 mL ethanol. After mixing and centrifugation (900 x g for 10 min at room temperature), the supernate was transferred to 2-mL vials. After the supernate was evaporated to dryness, sugars were derivatized to the pentaacetate derivative by using 75 µL acetic acid anhydride–pyridine (10:5 by vol) and reacting for 1.5 h at room temperature. After evaporation of the reagent, the derivatives were dissolved in 500 µL chloroform.

The ¹³C-¹²C ratio of glucose was determined by using gas chromatography–combustion isotope ratio mass spectrometry (GC-CIRMS) with a Delta S/GC instrument (Finnigan MAT, Bremen, Germany). The GC conditions were as follows: 2 µL of the chloroform solution was injected in the splitless mode onto a 25 m x 0.32 mm (0.2-µm film thickness) OV1701 column (CP Sil 19CB; Chrompack, Middelburg, Netherlands) installed in a Varian 3300 gas chromatograph. The oven temperature was programmed from 100°C (1 min) to 275°C (2 min) at a rate of 30°C/min. Helium was used as the carrier gas at a column head pressure of 138 Pa. Eluting compounds were combusted on-line in a platinum-catalyzed cupric oxide oxidation reactor operating at 800°C. Solvent and compounds eluting before 4 min were kept out of the reactor by applying a helium backflush gas flow. Water vapor was removed by nafion tubing (DuPont, Wilmington, DE) and the carbon dioxide pulses formed in the reactor were transferred to the IRMS through an open split interface.

The ¹³CO₂-¹²CO₂ ratio measured by IRMS was corrected for ¹⁷O abundance (19) and the final ¹³C-¹²C ratio was expressed as ¹³C_PDB (20). The ¹³C_PDB values of the glucose and starch substrates were determined by total combustion using an on-line coupled elemental analyzer.

Calculations
The ¹³C_PDB value was converted to the atom % (AP) value. The AP values after ingestion of substrate were corrected for the baseline abundance. The difference [atom % excess (APE)] was used for further calculations. For breath carbon dioxide, the APE value was related to the carbon dioxide excretion measured by indirect calorimetry to calculate the substrate-derived carbon dioxide exhalation rate. This exhalation rate was calculated as the percentage dose recovered (PDR)/h. Total recovery after 6 h was calculated as the area under the PDR/h-time curve and expressed as the cumulative percentage dose recovered (cPDR). The ¹³C_PDB of the plasma glucose pentaacetate derivative was converted to the ¹³C AP. The concentration of glucose derived from the [¹³C]starch (exogenous glucose) in plasma was calculated as follows:

RESULTS  
The 2-h AUC for total glucose concentration in serum derived from the intake of DCS, Hylon VII, and Novelose 330 related to the same measure after glucose intake, leads to the glycemic index values given in Table 1. High SDs for the obtained mean values were found under these experimental conditions.

View this table:
TABLE 1.. Characterization of digestibility of starch¹  
The ¹³C-¹²C ratio in serum glucose is presented in Figure 1A and the total glucose response in serum after consumption of 40 g glucose is presented in Figure 1B. The combination of both values per time point resulted in a time curve of exogenous glucose response (Figure 1B). In Figure 2, the total glucose response in serum after consumption of DCS, Hylon VII, and Novelose 330 is compared with the exogenous glucose response in serum in the same experiments. It is clear that the differences between DCS and Hylon VII and Novelose can be documented better by using the serum exogenous glucose response. Because of its high digestibility, DCS gave a more pronounced rise in exogenous glucose concentration in serum than did Hylon VII or Novelose.

View larger version (14K):
FIGURE 1. . After administration of 40 g glucose to 7 healthy volunteers, the mean (±SD) ¹³C-¹²C ratio in serum glucose was analyzed by using gas chromatography–combustion isotope ratio mass spectrometry (A; -); the mean (±SD) total glucose concentration was measured in the same samples (B; -). These 2 values were used to calculate the exogenous glucose concentration in serum (B; -). The calculation is described in Methods.

 

View larger version (15K):
FIGURE 2. . Mean (±SD) glucose response in serum after consumption of 40 g glucose (-), digestible cornstarch (-), Hylon VII (type 2 resistant starch; -), and Novelose 330 (type 3 resistant starch; -) in 7 healthy volunteers. Both the total glucose response (A) and the exogenous glucose response (B) are shown.

 
In Figure 3, the cumulative exogenous glucose curve after intake of glucose and the starches is illustrated. In the first period of the curve (t < 60 min), effects of variations in gastric emptying play a relatively major role. In the latter part of the curve (t > 3 h), postabsorptive events have more influence. Therefore, we decided to compare the cumulative exogenous glucose response in serum after glucose and starch intake at t = 2 h, which is the same period used in the classic glycemic index calculation method of Jenkins et al. When these calculations were made, ¹³C glycemic index values were derived (Table 1). These values more closely resemble the amylose-amylopectin ratio of starch than do those obtained by the classic glycemic index method (Table 1).

View larger version (18K):
FIGURE 3. . Mean (±SD) cumulative exogenous glucose response in serum (area under the concentration-time curve; AUC) after consumption of 40 g glucose (-) or starch in 7 healthy volunteers: digestible cornstarch (-), Hylon VII (type 2 resistant starch, -), and Novelose 330 (type 3 resistant starch, -).

 
The total carbon dioxide excretion in breath after administration of 40 g glucose and the ¹³C-¹²C ratio in breath samples determined by IRMS analysis are shown in Figure 4A. By combining these data, the ¹³CO₂ excretion rate (Figure 4B) and 6-h cPDR (Figure 4C) were calculated. The results when this method was applied after consumption of DCS, Hylon VII, and Novelose 330 are shown in Figure 5. Again, a more pronounced response was found after consumption of the well-digested DCS than after the resistant starch sources. Simultaneous measurement of hydrogen in breath indicated that in the 6-h period, no increase in breath hydrogen occurred, which suggests that the starch was not fermented in this period (data not shown). From this we concluded that the measured ¹³CO₂ was derived from digested starch and not from fermented starch.

View larger version (57K):
FIGURE 4. . In breath, the total carbon dioxide excretion (-) and mean (±SD) percentage ¹³C (-) was analyzed after consumption of 40 g glucose in 7 healthy volunteers (A). From these data, ¹³C excretion rate could be calculated (B) and expressed as a cumulative curve (C).

 

View larger version (20K):
FIGURE 5. . Mean (±SD) excretion rate of ¹³C in breath after consumption of 40 g glucose (-), digestible cornstarch (-), Hylon VII (type 2 resistant starch; -), and Novelose 330 (type 3 resistant starch; -) in 7 healthy volunteers.

 
To quantify the differences in the digestion process, the ¹³CO₂ response to corn-derived glucose was used. In Figure 6, the ¹³C cumulative excretion in breath after consumption of glucose, DCS, Hylon VII, and Novelose 330 is shown. Roughly, the difference in response to corn-derived glucose and cornstarch reflects intestinal hydrolysis. Assuming that glucose is absorbed completely from the small intestine, the partial digestion of starch can be derived by comparing the 6-h cumulative ¹³C excretion after corn-derived glucose and cornstarch consumption. These data were calculated for DCS, Hylon VII, and Novelose 330 and are presented in Table 1 as the ¹³CO₂ glycemic index. In this case, a 6-h period was used to minimize the problem of variable intermediate storage of glucose in tissues before oxidation.

View larger version (19K):
FIGURE 6. . Cumulative mean (±SD) excretion of ¹³C in breath after administration of 40 g glucose (-), digestible cornstarch (-), Hylon VII (type 2 resistant starch; -), and Novelose 330 (type 3 resistant starch; -) in 7 healthy volunteers.

 
Both techniques of measuring the exogenous glucose concentration in serum and measuring the ¹³CO₂ excretion in breath are indirect reflections of intestinal digestion. In an attempt to optimize these variables, the kinetics of both techniques were compared. In Figure 7, the values for the exogenous glucose curve in serum and the ¹³CO₂ excretion rate in breath after consumption of glucose, DCS, and Hylon VII are compared. As can be observed, there is a high degree of similarity; the differences are due to a lag phase in the breath excretion, as would be expected. Furthermore, the differences are larger for the glucose and DCS curves than for the Hylon VII curves.

View larger version (55K):
FIGURE 7. . Comparison of exogenous glucose response in serum (-) and ¹³CO₂ excretion rate in breath (-) after consumption of glucose (A), digestible cornstarch (B), and Hylon VII (type 2 resistant cornstarch; C).

 

DISCUSSION  
Short-chain fatty acids fulfill important biological functions in the colon (21). Butyrate has a distinct role in epithelial cell proliferation and differentiation (22, 23). It is not known what the physiologic range of butyrate production is in humans, but because of changes in food consumption patterns, daily intakes of substrates for short-chain fatty acids are steadily decreasing. A way to increase the supply of butyrate in the colon could be the consumption of resistant starch sources (24–26). To understand and evaluate the biological effects of consumption of resistant starch, we need to know first to what degree the resistant starch is digested in the small intestine. Information on digestibility is difficult to obtain because of the inaccessibility of the intestinal tract in humans. Several techniques to quantify the amount of starch that is resistant to the action of small intestinal enzymes have been used. The shortcomings of these techniques were discussed elsewhere (3, 4, 9, 11). We based our studies on the pioneering experiments of Normand et al (13), who described the use of ¹³C-enriched starch for in vivo starch digestion studies. The hydrolysis product, [¹³C]glucose in serum, was analyzed by GC-CIRMS, and the metabolic end product, ¹³CO₂ in breath, was analyzed by IRMS. We followed the same approach using cornstarch products, which are naturally enriched in ¹³C.

We calculated a ¹³C glycemic index based on comparison with the response to exogenous glucose. This ¹³C glycemic index was compared with the classic glycemic index; we found that the SDs for the obtained mean values were smaller with the ¹³C glycemic index than with the classic glycemic index. This may be because the ¹³C-glycemic-index method refers to exogenous glucose only and the classic glycemic index test cannot discriminate between endogenous and exogenous glucose. Furthermore, the values resemble the percentage of amylopectin in the starch more closely. A clear difference between the response to DCS, which is a highly digestible starch, and that to resistant starch can be found. DCS was digested an average of 80% and resistant starch 50%. This last value agrees with the data of Champ et al (11), who used retrograded amylose starch. [¹³C]glucose is ultimately oxidized to ¹³CO₂ and excreted via the breath. By measuring the excretion rate and cumulative excretion, the intestinal digestion of starch and the subsequent absorption of glucose can also be studied. This approach was used by Hiele et al (14, 27) to study the digestion of raw crystalline cornstarch. Also, with ¹³CO₂ breath tests [¹³C]glucose is used as a reference substrate. The assumption is that [¹³C]glucose derived from a glucose or a starch load undergoes similar metabolic events.

By using this approach, data about intestinal digestion of DCS, Hylon VII, and Novelose 330 were obtained (Table 1). These data led to the conclusion that DCS is effectively digested and resistant starch is digested 50%. No difference between Novelose 330 and Hylon VII was observed, despite more processing of the first compound. Many intermediate factors, such as differences in insulin response, tissue glucose clearance, and variable glucose oxidation rates, explain the intra- and interindividual differences in the outcome. The overall similarity in final outcomes via the [¹³C]glucose method and the ¹³CO₂ method is surprising and indicates that the rate-limiting step in the total process of converting raw starch to carbon dioxide is the hydrolysis step.

To optimize the proposed method for use in future studies to characterize factors involved in small intestinal digestion, the differences between the data obtained with exogenous glucose concentration in serum and ¹³CO₂ excretion in breath were highlighted. The results indicate that the differences are smaller when the starch is digested slower. Obviously, with a rapid influx of glucose from the small intestine, the insulin response will be higher and, consequently, there will be storage in tissues. This relation between exogenous glucose, endogenous glucose, and insulin will be studied in detail in future experiments. The results of the present experiment indicate that the kinetics of starch digestion can be derived from the [¹³C]glucose response in serum as well as the ¹³CO₂ excretion rate in breath and that these measurements can be used as a first screening of the digestibility of a starch source.

REFERENCES  

1. Englyst HN, Macfarlane GT. Breakdown of resistant and readily digestible starch by human gut bacteria. J Sci Food Agric 1986; 37:699–706.
2. Asp NG, van Amelsvoort JMM, Hautvast JGAJ. Nutritional implications of resistant starch. Nutr Res Rev 1996;9:1–31.
3. Cummings JH, Englyst HN. Measurement of starch fermentation in the human large intestine. Can J Physiol Pharmacol 1991;69:121–9.
4. Strocchi A, Levitt MD. Measurement of starch absorption in humans. Can J Physiol Pharmacol 1991;69:108–10.
5. Englyst HN, Kingman SM, Cummings JH. Classification and measurement of nutritionally important starch fractions. Eur J Clin Nutr 1992;46(suppl):S33–50.
6. Muir JG, O'Dea K. Measurement of resistant starch: factors affecting the amount of starch escaping digestion in vitro. Am J Clin Nutr 1992;56:123–7.
7. Champ M. Determination of resistant starch in foods and food products: interlaboratory study. Eur J Clin Nutr 1992;46(suppl):S51–62.
8. Muir JG, O'Dea K. Validation of an in vitro assay for predicting the amount of starch that escapes digestion in the small intestine of humans. Am J Clin Nutr 1993;57:540–6.
9. Englyst HN, Kingman SM, Hudson GJ, Cummings JH. Measurement of resistant starch in vitro and in vivo. Br J Nutr 1996;75:749–55.
10. Silvester KR, Englyst HN, Cummings JH. Ileal recovery of starch from whole diets containing resistant starch measured in vitro and fermentation of ileal effluent. Am J Clin Nutr 1995;62:403–11.
11. Champ MM, Molis C, Flourie B, et al. Small-intestinal digestion of partially resistant cornstarch in healthy subjects. Am J Clin Nutr 1998;68:705–10.
12. Hatch MD, Slack CR. Photosynthetic CO₂-fixation pathways. Annu Rev Plant Physiol 1970;21:141–62.
13. Normand S, Pachiaudi C, Khalfallah Y, Guilluy R, Mornex R, Riou JP. ¹³C appearance in plasma glucose and breath CO₂ during feeding with naturally ¹³C-enriched starchy food in normal humans. Am J Clin Nutr 1992;55:430–5.
14. Hiele M, Ghoos Y, Rutgeerts P, Vantrappen G, de Buyser K. ¹³CO₂ breath test to measure the hydrolysis of various starch formulations in healthy subjects. Gut 1990;31:175–8.
15. Dewit O, Prentice A, Coward WA, Weaver LT. Starch digestion in young children with cystic fibrosis measured using a ¹³C breath test. Pediatr Res 1992;32:45–9.
16. Weaver LT, Dibba B, Sonko B, Bohane TD, Hoare S. Measurement of starch digestion of naturally ¹³C-enriched weaning foods, before and after partial digestion with amylase-rich flour, using a ¹³C breath test. Br J Nutr 1995;74:531–7.
17. Jenkins DJ, Wolever TM, Taylor RH, et al. Glycemic index of foods: a physiological basis for carbohydrate exchange. Am J Clin Nutr 1981;34:362–6.
18. Koetse HA, Stellaard F, Bijleveld CMA, et al. Non invasive detection of low intestinal lactase activity in children by use of a combined ¹³CO₂/H₂ breath test. Scand J Gastroenterol 1999;34:35–40.
19. Mook WG, Grootes PM. The measuring procedure and corrections for the high-precision mass-spectrometric analysis of isotopic abundance ratios, especially referring to carbon, oxygen and nitrogen. Int J Mass Spectrom Ion Phys 1973;12:273–98.
20. Craig H. Isotopic standards for carbon and oxygen and correction factors for mass-spectrometric analysis of carbon dioxide. Geochim Cosmochim Acta 1957;12:133–49.
21. Kim YI. Short-chain fatty acids in ulcerative colitis. Nutr Rev 1998;56:17–24.
22. McIntyre A, Gibson PR, Young GP. Butyrate production from dietary fibre and protection against large bowel cancer in a rat model. Gut 1993;34:386–91.
23. Heerdt BG, Houston MA, Augenlicht LH. Potentiation by specific short-chain fatty acids of differentiation and apoptosis in human colonic carcinoma cell lines. Cancer Res 1994;54:3288–93.
24. Weaver GA, Krause JA, Miller TL, Wolin MJ. Cornstarch fermentation by the colonic microbial community yields more butyrate than does cabbage fiber fermentation; cornstarch fermentation rates correlate negatively with methanogenesis. Am J Clin Nutr 1992;55:70–7.
25. Mathers JC, Dawson LD. Large bowel fermentation in rats eating processed potatoes. Br J Nutr 1991;66:313–29.
26. Scheppach W, Fabian C, Sachs M, Kasper H. The effect of starch malabsorption on fecal short-chain fatty acid excretion in man. Scand J Gastroenterol 1988;23:755–9.
27. Hiele M, Ghoos Y, Rutgeerts P, Vantrappen G. Starch digestion in normal subjects and patients with pancreatic disease, using a ¹³CO₂ breath test. Gastroenterology 1989;96:503–9.