Digestibility: amino acids

The commonly used term “amino acid digestibility” does not of course refer to the degradation of amino acids but is shorthand for the proportion of consumed amino acid that is absorbed.

It is worth emphasizing at the outset that digestibility is not a fixed attribute of a food but reflects an interaction between the food and the person or animal eating it. Of course, the composition of a given food is subject to many sources of variation from one sample to the next, but in a given sample chemical composition can - in theory at least - be determined with high precision. Its digestibility however varies according to how and to whom it is fed, so that the digestibility of a food (or rather of a nutrient within that food) cannot be assigned a unique value relevant to all species or all individuals of a species.

For amino acids the current method of choice is ileal digestibility⁽Reference Stein, Sève and Fuller³⁾. The road to this consensus has been a long one and this is not the place to rehearse all of the evidence that has led there⁽Reference Stein, Sève and Fuller^3–Reference Darragh and Hodgkinson⁵⁾ but the reasons for this choice stem from two seminal observations. The first observation⁽Reference Mason and Palmer⁶⁾ was that most faecal nitrogen is in the form of microbial protein. Mason et al. ⁽Reference Mason, Just and Bech-Andersen⁷⁾ estimated from the faecal excretion of diaminopimelic acid (DAPA) that as much as 90 % of the faecal N of pigs could be of bacterial origin. Subsequent studies using a variety of microbial markers have confirmed this observation. Consequently, the amino acid composition of faeces tends to be closer to that of microbial protein than to that of undigested food residues, so the amino acid composition of faeces varies little with diet. It was concluded that undigested food residues reaching the large intestine are largely degraded by microbial activity during their relatively long residence, their nitrogen being either absorbed or converted into microbial biomass with an amino acid profile more or less independent of their initial composition.

The second observation was that although nitrogen can be absorbed from the large intestine, there is (except perhaps in neonates) little absorption of intact amino acids. This was first demonstrated by Zebrowska⁽Reference Zebrowska⁸⁾, who infused hydrolyzed casein into the caecum of pigs. There was very little increase in faecal nitrogen but some 83–90 % of the infused N was excreted in the urine with little if any improvement in N retention, in contrast to the substantial response to an oral casein supplement. Other studies, with infusions of protein or amino acids, have confirmed this (see⁽Reference Fuller and Reeds⁹⁾). This means that most of the carbon skeletons of indispensable amino acids entering the large intestine from the ileum are irreversibly lost, either through microbial metabolism or excretion in the faeces, although their nitrogen may be absorbed and used, as will be discussed later. It is worth emphasizing that the appearance of ¹⁵N-labeled proteins in the blood after caecal administration of ¹⁵N-labeled proteins (e.g.⁽Reference Heine, Wutzke and Richter¹⁰⁾) is not evidence of the absorption of amino acids: most amino acids in the body can acquire ¹⁵N by transamination, as seen in the extensive ¹⁵N labeling of body protein after giving ¹⁵NH₄Cl (e.g.⁽Reference Patterson, Carraro and Klein^11, Reference Metges, Petzke and El-Khoury¹²⁾). The infusion into the caecum of growing pigs of a single indispensable amino acid that was deficient in the diet has been shown to be of little or no benefit⁽Reference Wünsche, Hennig and Meinl^13–Reference Darragh, Cranwell and Moughan¹⁵⁾. As pointed out elsewhere⁽Reference Fuller and Reeds⁹⁾, with infusion into the caecum of protein, rather than single amino acids, however, there has generally been some positive change in nitrogen retention⁽Reference Zebrowska^8, Reference Zebrowska^16, Reference Gargallo and Zimmerman¹⁷⁾. Because most of the N infused into the caecum was absorbed, this may be attributable to the addition of non-specific nitrogen, in the same way that dietary supplements of non-essential N such as urea or ammonium salts can improve nitrogen balance⁽Reference Swendseid, Harris and Tuttle^18–Reference Kofrányi and Jekat²⁰⁾. An increased supply of non-amino N can be expected to increase urea recycling and microbial amino acid synthesis in the small intestine, where the microbial amino acids seem to be mostly absorbed⁽Reference Torrallardona, Harris and Fuller²¹⁾, which may be of nutritional benefit⁽Reference Raj, Dileep and Vaz²²⁾. Taken together therefore the evidence suggests that most of the carbon skeletons of amino acids entering the large intestine from the ileum are irreversibly lost, either through microbial metabolism or excretion in the faeces, although their nitrogen may be absorbed and used.

In consequence of these observations it is widely agreed that estimates of the amino acids absorbed from the diet would best be derived from measurement of the flow of amino acids leaving the small intestine; that is, ileal digestibility. However, not all the amino acids leaving the ileum are of immediate dietary origin: some are the remnants of endogenous secretions and cellular material: their loss represents part of the requirement and must be deducted from the amino acid flow in order to estimate the contribution of unabsorbed amino acids from the diet. This is the correction of apparent to true digestibility.

For protein as a whole, however, because nitrogen absorbed in forms other than as amino acids can contribute to the nitrogen economy, the absorption of nitrogen over the whole digestive tract is the more appropriate measure. This simply requires the collection and analysis of faeces, which is straightforward; it also requires correction for endogenous losses.

The required measures are therefore of nitrogen and amino acid intakes, of ileal amino acid outflow and of faecal nitrogen loss, together with appropriate values for the endogenous components, which will be considered in more detail in later papers⁽Reference Moughan and Rutherfurd^23, Reference Columbus and De Lange²⁴⁾.

The measurement of ileal amino acid outflow

Methods for measuring ileal amino acid outflow have been reviewed elsewhere⁽Reference Fuller^25–Reference Moughan²⁸⁾ and only a brief synopsis will be given here. The methods can be divided into two classes: those in which the whole outflow is collected and those in which only a sample is collected, which is related to the whole flow by use of an indigestible marker. To collect the whole flow requires an ileostomy or ileo-rectal anastomosis and the continuous collection must be sufficiently long for day-to-day variation to be minimized. Whilst obviating the need for a marker, each of these approaches brings its own potential problems. Removal of the ileal digesta means that the potential for recycling nitrogen from the large intestine is lost, with possible consequences for the nitrogen transactions of the small intestine. Although this objection may be overcome by returning the digesta to the large intestine after sampling via a second cannula⁽Reference Darcy, Laplace and Villiers²⁹⁾, the disturbance of normal gut motility by the transection of the gut may in turn alter the digestive processes. A further concern is that, following an ileostomy or ileo-rectal anastomosis, the microbial population and the morphology of the ileum undergo changes that may alter the nature and extent of digestion there so that they more closely resemble those of the caecum and colon⁽Reference Gorbach, Nahas and Weinstein^30, Reference Köhler, Mosenthin and Verstegen³¹⁾. This may be more relevant with an ileostomy than with an ileorectal anastomosis as Hennig et al. ⁽Reference Hennig, Metges and Berk³²⁾ reported little change with time in ileal microbial numbers during a 5-month period after ileo-rectal anastomosis in pigs. A further and perhaps decisive concern is the welfare of animals with an ileo-rectal anastomosis.

Digesta collection involving indigestible markers allows shorter and simpler procedures. At its simplest, animals can be killed some time after a test meal and the digesta in the terminal ileum can be removed and the concentrations of amino acids and of marker estimated. Although simple, this approach has several practical problems. First, each animal can be used only once, so that variation between animals cannot be reduced as it can by repeated within-animal measurements. Second, especially with small animals, there may be insufficient digesta at the very end of the ileum for the necessary analyses. Collecting digesta from a longer segment of the ileum means that some will be less completely digested.

Commonly used markers include chromic oxide, titanium oxide and acid-insoluble ash. Other heavy metals and rare earth elements have also been used. Dual-phase markers have also been advocated on the premise that amino acids are contained in both liquid and solid phases of the digesta and both need to be represented. Although there have been some comparisons of markers there does not seem to be any consensus as to the relative merits of the various substances.

All methods involving indigestible markers involve the assumption that the marker is distributed in the digesta in the same way as the amino acid (or other nutrient) being measured. This does not necessarily require that the digesta be homogeneous but that the proportion of the intake of marker in the sample be the same as the proportion of the total amino acid flow that is in that sample. This in turn requires the assumption that all the marker fed should reach the point of sampling. Although some investigators have reported virtually complete recovery⁽Reference Fairweather-Tait, Minihane and Eagles³⁶⁾ others have reported recovering significantly less than 100 %⁽Reference Köhler, Huisman and den Hartog^37–Reference Mroz, Bakker and Jongbloed³⁹⁾, especially of chromic oxide. Marker recovery may well vary amongst marker substances and methods of preparation and administration but complete recovery of the chosen marker is often not tested.

To collect digesta repeatedly and without interrupting normal digesta flow various cannulas have been developed, the simplest being a T-cannula in the terminal ileum. When the T-branch is opened a variable proportion of the flow is diverted to the outside where it can immediately be chilled until analyzed. For most diets this works well but, especially with fibrous materials, there may be a fractionation of the digesta so that the material exiting the cannula is no longer representative of the whole flow. To overcome this problem a post-valvular cannula has been developed⁽Reference Van Leeuwen, van Kleef and van Kempen^33, Reference van Leeuwen, van Kleef and van Kempen³⁴⁾. This is placed just distal to the ileocaecal junction but so arranged as to divert, while the sample is being collected, practically the whole outflow through the ileo-caecal valve to the exterior. The advantages of this technique now seem to make it the method of choice with pigs. Whatever method is used to obtain digesta it is essential that digesta are collected as soon as they emerge and further bacterial activity prevented: failure to do so results in exaggerated NSP digestibility and presumably other fermentative changes⁽Reference Canibe and Bach Knudsen³⁵⁾.

Apparent, true and standardized ileal digestibility

Subtracting ileal amino acid outflow from amino acid intake gives apparent digestibility. However, because of the contribution of endogenous amino acids to the ileal outflow - a contribution that is neither constant nor proportional to the protein intake - the value obtained depends on the way the food has been given. In particular, the test protein must be included in the diet above a certain threshold concentration to obtain plateau values of apparent digestibility⁽Reference Fan, Sauer and Hardin^40–Reference Eklund, Mosenthin and Piepho⁴²⁾. However, to obtain digestibility values for individual foods that can be used additively to predict the digestibility of a complete diet, apparent digestibility values must be corrected appropriately for the endogenous losses. For this purpose endogenous losses are considered as consisting of three components, a basal loss that is independent of the diet and which forms part of the animal's requirement, an additional loss that is induced by giving protein (or peptides, or amino acids) per se (regardless of their source or digestibility) and a further loss which can be accounted as a charge against the particular food, reducing the proportion that is available to meet metabolic requirements. The time-honoured term true digestibility refers to the value corrected for all endogenous losses: the more practical measure is the one corrected only for basal losses, that is, standardized digestibility⁽Reference Stein, Sève and Fuller³⁾. These issues, as well as endogenous losses and their measurement form the subject of other contributions to this symposium⁽Reference Moughan and Rutherfurd^23, Reference Columbus and De Lange²⁴⁾ and will not be discussed further here.

Measurement of ileal outflow in human subjects

Clearly, most of the methods used to estimate ileal amino acid flows in animals cannot be considered with human subjects. The exception is the use of volunteers who have an ileostomy. This allowed the first direct comparison between measurements in man and those in animals given the same diets⁽Reference Rowan, Moughan and Wilson⁴³⁾. Although, as mentioned above, the physiology and microbiology of the ileum are altered following ileostomy, the quantitative effects of these changes on protein digestion and ileal amino acid outflow are not yet known. For example⁽Reference Rowan, Moughan and Wilson⁴³⁾, the ileal digestibility of pectin and hemicellulose in ileostomates was 0·87 and 0·45 and the daily output of DAPA in ileostomy fluid was 31 mg, about 40 % of that in faeces, which suggests a distribution of fermentation between the small and large intestine similar to that in the pig. However, we do not know what these values would be in intact subjects.

An alternative approach to measuring the flow and composition of ileal digesta in human subjects has been made by the infusion of a marker and withdrawal of samples through a naso-intestinal tube⁽Reference Mahé, Huneau and Marteau⁴⁴⁾. This allowed the digestibility of the same proteins to be compared in pigs and intact human subjects⁽Reference Deglaire, Bos and Tomé⁴⁵⁾. Those comparisons were of purified proteins: the very small diameter of the sampling tube (1·5 mm) probably limits the application of this approach to such materials, or possibly to very finely ground foods. In any case, this method is too demanding for routine use and is likely to be restricted to validating the application to humans of digestibility data obtained by other means, either another animal species or an in vitro system (e.g., ⁽Reference Boisen and Eggum^46, Reference McDonough, Sarwar and Steinke⁴⁷⁾). The approach of Le Gall et al. ⁽Reference Le Gall, Quillien and Guéguen^48, Reference Le Gall, Quillien and Seve⁴⁹⁾, in which the peptide residues of undigested dietary proteins and endogenous proteins in ileal digesta were identified by immuno-blotting and mass spectrometry might also be used to examine the extent of digestion of specific proteins in man.

Surrogates for humans

The principal candidates for providing digestibility data applicable to man are other animals and in vitro systems. The primary requirement of course is simply that the proposed system should provide digestibility values that are either the same as those obtained in man, or that can be transformed numerically by a consistent formula. However, given the paucity of data obtained with human subjects a more feasible starting point is to set out certain contributory requirements.

1. (1) The digestive system should be similar to that of man.

2. (2) The system should accept diets as consumed, and not require additional processing.

3. (3) Diets should be completely consumed, with no selection.

Animals that are commonly used in digestibility studies are the rat, the chicken and the pig.

The first requisite probably excludes birds, which have a gizzard and do not express lactase activity; they also have a shorter retention time of food in the GIT. It probably also excludes rats unless coprophagy is prevented. In addition, to prevent food selection by rats diets must be finely ground or otherwise processed before feeding; processes that may alter digestibility⁽Reference Lahaye, Ganier and Thibault⁵⁰⁾. The digestive system of the pig differs from that of man principally in having a large caecum; however, this is not relevant to ileal digestibility. Pigs do not normally practice coprophagy and, true to their reputation, readily eat unprocessed diets without selection.

Possible limitations of ileal digestibility: the role of the upper intestinal microbiota

As discussed earlier, the reason for rejecting faeces in favour of ileal digesta as the basis for assessing amino acid digestibility is the modifying effect of the large intestinal microbiota on the amino acid composition of the digesta. The upper digestive tract, however, has its own microbiota. Only since the development of culture-independent methods have the extent and diversity of this population been apparent⁽Reference Leser, Amenuvor and Jensen^51, Reference Booijink, El-Aidy and Rajili-Stojanovi⁵²⁾. Many of the species have never been cultured and we do not yet have a clear picture of the impact of this population on the nitrogen transactions that are of interest to us in terms of digestion and absorption of amino acids. However, there is evidence that its impact, in pigs at least, is considerable. It was estimated, from the concentration of DAPA in the ileal effluent of pigs, that the microbial biomass leaving the ileum was a quarter of that in the faeces⁽Reference Mason, Just and Bech-Andersen⁷⁾. Other estimates in pigs show that up to half the nitrogen leaving the ileum may be in microbial biomass⁽Reference Poppe, Meyer and Bennke^53–Reference Miner-Williams, Moughan and Fuller⁵⁵⁾. Studies of microbial metabolic activity, measured as adenylate energy charge, are as high in the ileum as in the large intestine in pigs⁽Reference Bach Knudsen, Jensen and Andersen⁵⁶⁾. Related to this microbial activity, substantial digestion of various non-starch polysaccharides has been shown to occur proximal to the ileo-caecal junction⁽Reference Millard and Chesson^58–Reference Bach-Knudsen and Canibe⁶⁰⁾, providing an energy source for microbial amino acid synthesis. It is important to bear in mind that these observations were made in cannulated animals in which there may have been some modification of the normal microbial population and its environment.

Although there have been several studies of the microbiology of human ileostomy fluid⁽Reference Booijink, El-Aidy and Rajili-Stojanovi^52, Reference Ruseler-van-Embden, Kool and van Lieshout^61, Reference Fuller, Milne and Harris⁶²⁾ there is understandably rather little information about the normal human ileal microbiota. Ileal digesta have been reported to contain 10⁷-10⁸ bacteria/mL⁽Reference Hao and Lee^63–Reference Tannock⁶⁵⁾, two to three orders of magnitude less than in faeces. More recent analyses⁽Reference Hayashi, Takahashi and Nishi^66, Reference Wang, Heazlewood and Krause⁶⁷⁾ using nucleic acid-based methods have revealed the diversity of organisms in the ileum, the similarities and differences between ileal and colonic flora, and considerable differences between individuals, but each study included samples from only two or three individuals, taken during surgery or at autopsy. Very recent reports based on ileal digesta sampled by naso-intestinal tube provide a clearer picture of the microbiota of the normal ileum⁽Reference van den Bogert, de Vos and Zoetendal^68, Reference Zoetendal, Raes and van den Bogert⁶⁹⁾.

What are the consequences of this microbial activity on the fate of dietary amino acids (and of amino acids in endogenous proteins that may potentially be reabsorbed)? What is the impact of microbial amino acid synthesis in the upper GI tract on estimates of ileal amino acid digestibility?

The gastrointestinal microbiota can both degrade and synthesize amino acids and, given the heterogeneity of the intra-luminal environment, both may occur simultaneously. When Nesheim & Carpenter⁽Reference Nesheim and Carpenter⁷⁰⁾ pointed out many years ago that microbial degradation of amino acids in the gut would lead to a spurious over-estimate of digestibility, they were referring to faecal nitrogen but the same is true for ileal amino acid digestibility. Equally, a net synthesis of amino acids by the microbiota would lead to an under-estimate of digestibility. Both possibilities need to be considered.

Integrative bioavailability assays

As mentioned earlier digestibility is not the only determinant of bioavailability, although it is usually the most important. Other factors, such as the chemical availability of individual amino acids, notably lysine (and other amino acids) in heat-damaged proteins, can be estimated separately⁽Reference Carpenter^91, Reference Moughan and Rutherfurd⁹²⁾. Alternatively, bioavailability may be estimated by one of several bioassays that integrate the several factors that limit the extent to which a dietary amino acid is absorbed and available for metabolism. Using a response that is sensitive to amino acid supply these approaches compare individual foods to pure amino acids, usually in the form of a slope-ratio assay. A suitable response, such as growth rate or body protein accretion⁽Reference Batterham⁹³⁾ or AA oxidation⁽Reference Moehn, Bertolo and Pencharz⁹⁴⁾, of animals fed the test ingredient is related to the AA intake, and the slope of the regression line is compared with that obtained with the pure amino acid. All diets used in this assay must be first-limiting in this AA, and all dietary levels of the AA need to be below the requirement. It is assumed that the response to graded AA intake levels is linear and determined only by the amino acid in question. However, despite the attractions of a method that integrates all the factors that may limit the utilization of dietary amino acids, the approach has the practical disadvantage that a separate assay is required for each amino acid, using a series of diets in which only that amino acid is limiting. It also tends to confound bioavailability with issues of nutrient imbalance, complicating the interpretation of the results. Furthermore, these assays are laborious and costly and, compared with ileal digestibility measurements, the estimated availability values generally have higher standard errors of determination.