Camp. &o&em. Physiol. Vol. 97A, No. I, pp. 5743, Printed in Great Britain
1990
0300-9629/90 $3.00 + 0.00 0 1990 Pergamon Press pie
EFFECTS OF GENETIC SELECTION ON GROWTH RATE AND INTESTINAL STRUCTURE IN THE DOMESTIC FOWL (GALLUS DOMESTICU,!2) M.W.
SMITH,*M. A. MITCHELL?and M. A. PEACOCK* *AFRC Institute of Animal Physiology and Genetics Research, Babraham, Cambridge CB2 4AT, UK. Telephone: 0223-832312;and tRosiin, Midlo~ian EH25 9PS, UK. Telephone: 031-440-2726
Abstract-I.
Pieces of small intestine taken from chickens subjected previously to continuous selection, relaxed selection or no selection for rapid growth were used to estimate villus surface area and microvillus development to determine what effects genetic selection might have on factors controlling intestinal function. 2. Crypt size and the rates at which enterocytes migrated out of crypts were also measured, after injection of tritiated thymidine, to determine the time course of microvillus elongation. 3. Differences in growth rates measured between highly selected, relaxed selected or unselected birds were found to be correlated with parallel changes in villus surface area. Selection for growth did not change the density, dimensions or pattern of development of enterocyte microviili. Mi~rovilli did, however, produce a maximal 20-fold increase in villus surface area under all conditions. 4. Crypt size and enterocyte migration rates did not vary significantly between tissue taken from unselected and relaxed selected chickens. Tissue taken from highly selected birds had a crypt size and enterocyte migration rate 40% higher than values found for the other two groups of chickens. 5. The possibility that early genetic selection increased growth potential by uncoupling diet-induced changes on crypt hyperplasia from secondary effects on villus structure, and that later selection increased growth potential by increasing appetite, is discussed.
MATERIALS AND METHODS
Changes in enterocyte function and villus surface area occurring during the lifetime of the animal or, more permanently, through chance mutation favouring survival, together account for the ability of the small intestine to modify nutrient capture and energy metabolism (James et al., 1988; Tivey and Smith, 1989). Microvilli present on the enterocyte surface further amplify the area of villus available for the digestion and absorption of nutrients (Wilson, 1962). Changes at this level have also been found to be linearly related to the metabolic live mass of different mammalian species (Ferraris et al., 1989). The present situation is therefore one where interest in determining both villus and microvillus structure is increasing in studies designed to identify factors controlljng intestinal function. Neglected so far in this type of consideration is work involving chickens, a species genetically selected by commercial breeders in order to improve growth performance. Present work uses chickens subjected previously to different levels of genetic selection to test the hypothesis that accelerated weight gain might be associated with major changes in intestinal function. Results obtained showed vilius area, but not microvillus structure, to be directly related to chicken growth rates. The possibility that increases in villus area enable faster growing birds to sustain increased demands for nutrient assimilation is discussed.
Four-week-old female chickens used in the present experiments came from three lines exposed to different degrees of selection pressure for rapid growth rate and feed conversion efficiency. A modem commercial line was employed which has been continuously selected upon these criteria until the present day (designated highly selected). A second line had been selected for growth rate until 1972 and then selection was discontinued (relaxed selected line). A third brownleghorn line had never been selected for increased growth rate (J-line unselected). Newly hatched chickens from all three sources were originally maintained in brooders kept at an environmental temperature of 30°C. This temperature was gradually reduced to 22°C over a 2 week period after which birds were transferred to individual cages in climatic chambers, where they were kept under controlled conditions of tem~rature and fighting (2210S”C: 14 hr Ii~t~lO hr dark). -A commercial- broie; m&h was fed ad ~~~it~~ thoughout to chickens allowed free access to water. Weight gain in these animals was monitored by daily weighing during the week prior to the experiment.
Small intestines removed from chickens killed by cervical dislocation were measured and samples of tissue taken 25, 50 and 75% along the tract for further analysis. Initial preparation involved flushing ail tissue samples with phosphate buffered saline to remove intestinal contents before dividing each portion into three for later dete~ination of villus structure by image analysis, enterocyte migration rate by autoradiography and microvillus development by electron microscopy. The depth of individual crypts was also
M. W. SMITH el al.
58
measured by image analysis of haematoxylin and eosin stained sections of material prepared for autoradiography. The amount by which enterocyte microvilli increased villus surface area was assessed separately by measuring the diameter, height and number of microvilli present per standard length of tissue section by electron microscopy. Complete analysis of the above parameters was restricted to the most proximal region of the small intestine. Estimates of enterocyte migration rate, villus and crypt structure, were carried out at all three positions along the gut. Measurement
of villus surface area
Intact villi microdissected from tissue fixed previously in phosphate buffer containing 4% v/v glutaraldehyde and 2% w/v sucrose, using fine Monoject hypodermic needles (0.5 x 16 mm), were mounted on microscope slides in a 1: 1 mixture of glycerol and phosphate-buffered saline. These villi were then flattened by applying a I kg force to a 2 x 2 cm cover-slip and the edges of the cover-slip sealed with Tipp-Ex to preserve villus conformation and prevent dehydration. The height and surface area of these villi was then determined by transferring images through a TV camera attached to a Leitz microscope into a Magiscan 2A image analyser. Enhanced contrast of these images was provided by passing light from the microscope lamp through a Glen Spectra narrow band filter. Values of surface area obtained using the standard Magiscan MENU program were finally doubled, it being assumed that each villus consisted of two flat sheets bent to meet each other only in the region of the villus perimeter. Measurement
of enterocyte migration rate
Initial labelling of crypt cells and subsequent migration of labelled enterocytes onto villi was determined by injecting tritiated thymidine intravenously (37 kBq/g body wt) into 42 chickens (14 per group). Injected animals were then killed 2, 24, 48 or 72 hr later to obtain tissue samples for fixation in sucrose-glutaraldehyde buffer. Fixed tissue washed in phosphate buffered saline was then dehydrated and embedded in glycol-methacrylate for sectioning on a Reichert-Jung Autocut microtome. Cut sections were then coated with Ilford K2 photographic emulsion for later development and enhancement of Ag grain densities. Both exposure and enhancement times needed to prepare suitable autoradiographs were much longer than reported previously when injecting tritiated thymidine intraperitoneally (Smith and Brown, 1989). This latter route of injection, which probably gives a long tissue exposure to label, is to be preferred to the intravenous route of administration. Prepared autoradiographs were later stained with eosin and the positions of the leading edges of labelled cells determined by image analysis. The distance migrated by labelled cells was linearly related to the time after thymidine injection. Regression analysis of this data gave calculated migration rates in pmm/hr for the three groups of chickens. Elecrron microscopical measurement of microvillus slructure Samples of small intestine taken from the proximal region of unselected, relaxed selected and highly selected chickens were cut open, pinned out onto cork and fixed flat in phosphate buffer containing 4% v/v glutaraldehyde and 2% w/v sucrose. Fixed tissues were then cut into small strips for post-fixation in Verona1 buffer containing 1% w/v osmium tetroxide. Post-fixed samples were finally dehydrated, embedded in Araldite and sectioned for examination in a Philips 400 electron microscope. Only those areas of tissue showing appropriately sectioned microvilli were used for measurement of microvillus length and diameter. Individual positions on the crypt-villus axis chosen for measurement of microvillus structure were also recorded using the cryptvillus junction as a reference point. Microvillus densities in crypts, at the crypt-villus junction and at positions on villi where microvillus elongation was
complete were measured separately by counting the number of microvilli present in a standard length of sectioned tissue. These values were then used together with estimates of microvillus length and diameter to calculate microvillus surface area per villus. Logistic curve analysis of microvilhs elongation The method adopted for fitting logistic growth curves to obtain constants describing quantitative aspects of microvillus formation during enterocyte migration from crypt base to villus tip was identical to that described previously (Smith et al., 1984; Smith and Peacock, 1989). Briefly it consisted of fitting logistic growth curves having the form y = a + c/{ I + exp[ - b(x - m)]} to measurements of microvillus length obtained along the whole crypt-villus axis of unselected, relaxed selected and highly selected chickens and comparing these results with a single logistic curve obtained from fitting combined results derived from all chickens. In this formula, a is the calculated initial length of the microvillus, c is the microvillus growth that occurs subsequently, m is the position or time when each enterocyte half completes this growth phase and b is an exponential coefficient. Slaristical analysis Unpaired Student’s r-tests were used to assess the statistical significance of differences found between mean results f SEM. Standard errors associated with calculated logistic curve constants refer to the goodness of fit of curves to the experimental data rather than differences existing between individual chickens. The F test was used to assess whe*tier the error involved in fitting three separate growth curves to data obtained from the three groups of chickens was significantly different from that found when fitting a single growth curve to pooled data. Materials The [mefhyl-3 H] thymidine (I .5-2.2 TBq/mmol) came from Amersham International plc, Amersham, UK. All other reagents used were of AR Grade. RESULTS
Genetic selection for chicken growth rate Chickens classified as being unselected, relaxed selected or highly selected for rapid growth were found to weigh 186 f 5, 523 f 10 and 775 f 20 g 3 weeks after hatching (means f SEM; 21-23 birds per group). These weights had increased significantly to 249 rf: 6, 711 rf: 12 and 1048 + 29g 1 week later. Weight gains measured daily in chickens during the period 34 weeks after hatching were related linearly to age of chicken in all three groups (Fig. 1). Fitting regression lines to these data gave daily weight gains of 10.6 f 0.4, 30.1 + 1.7 and 43.3 f 2.2 g for unselected, relaxed selected and highly selected birds respectively. It is concluded from these initial findings that all three groups of chickens are growing normally at this stage in development and that differences in absolute growth rates are sufficiently great as to warrant further investigation.
Genetic selection for intestinal structure and enterocyte replacement
Changes in intestinal structure were tested for in the three lines of chickens chosen for study by measuring villus surface area, villus height, crypt depth and enterocyte migration rate as described in the text. Results obtained for villus surface area are summarized in Fig. 2.
Genetic effects on intestinal structure
59
0.3 r
I
22
I
I
I
I
I
23
24
25
25
2l
I
25
Fig. 1. Effect of genetic selection upon chicken growth rate. Chickens which were either unselected (A), relaxed selected (0) or highly selected (0) for rapid growth were allowed free access to food and water until 4 weeks of age when they were killed for experiment. Values give mean weight gains measured during the week prior to experiment (21-23 chickens per group). Villus surface area from unselected chickens was less than that determined in relaxed selected birds. Largest villus surface area occurted in chickens
highly selected for growth. These changes took place without altering the overall shape of villi which remained roughly rectangular throughout. This constancy of proportionality allowed one to fit a single curve to all experimental findings relating surface area (A: mm2) to villus height (H: mm) using the
A = 1.26H’.‘r. Values for mean surface area determined for the three groups of birds, together with other measurements of intestinal structure, are summarized in Table 1. Villi taken from birds where selection had been relaxed had twice the surface area of those determined in unselected birds (1.6 vs 0.8 mm2). This increase took place without change in crypt depth or enterocyte migration rate. A further 50% increase in
equation
4-
3-
A-
d 4
91
a
s
2-
3 >
l-
1.0 Villus
length
1.5
PO
(mm)
Fig. 2. ElTect of genetic selection upon the height and surface area of chicken intestinal villi. Microdissected villi from the proximal small intestine were analysed from chickens which were unselected (A), relaxed selected (0) or highly selected (0) for rapid growth. The fitted curve relating surface area (A) to villus height (If) has the form A = 1.269H’,“.
M. W. St+ttr~ et al.
60 Table I. Effect of genetic selection
Relaxed selected Highly
selected
and enterocyte
kinetics
in chicken
VillUS height (mm)
Villus area (mm’)
Crypt depth (mm)
Enterocyte migration rate (rmlhr)
0.85 f 0.04 (15:45) I .09 * 0.05 (15:45) I .3l k 0.06 (15:45)
0.78 ? 0.03 (15:45) I.61 k 0.08 (15:45) 2.27 + 0.18 (15:45)
0.20 &-0.01 (15:150) 0.19~0.01 (15:150) 0.27 f 0.01 (15:257)
IO.5 * 1.0 (14:282) 10.1 _+0.6 (14:194) 14.0 It 0.6 (13:257)
Chicken line Unselected
upon intestinal structure proximal small intestine
Measurements of villus height and area were obtained by image analysis of microdissected tissue taken from chickens injected previously with tritiated thymidine. Enterocyte migration rate and crypt depth measurements were carried out on eosin stained autoradiographs prepared from adjacent pieces of tissue as described in the text. Values give means + SEM. Numbers in brackets show the number of chickens used and villi analysed respectively.
villus surface was seen in birds highly selected for rapid growth (2.3 vs 1.6 mm2). This change was associated with significant increases in both crypt depth and enterocyte migration rate (P < 0.001 in both cases). Further morphometic and enterocyte kinetic measurements, carried out in mid-distal regions of the small intestine, showed enterocyte migration to fall to 40-50% of that measured in the proximal small intestine (5.4 vs 10.5; 6.9 vs 10.1 and 10.1 vs 14.0 pm/ hr for unselected, relaxed selected and highly selected birds). These changes were not associated with any obvious alteration in crypt depth (0.18 vs 0.20; 0.19 vs 0.19 and 0.24 vs 0.27 mm, respectively). It is concluded that selective line-dependent changes in intestinal structure and enterocyte migration do take place along the whole length of the small intestine.
I - 0.3
I
I
I
0
I
I
0.3 Distance
from
Changes in villus surface area are moreover roughly proportional to differences noted earlier in chicken growth rates (Table 1; Fig. 1). Genetic selection for microvillus
Microvilli are known to cause large increases in apparent villus surface area in mammalian small intestines (Wilson, 1962; Ferraris et al., 1989). It was therefore considered necessary to measure development of microvillus structure in these three types of bird to test whether selection pressure could also be detected at this level of analysis. Results obtained from carrying out these measurements are summarized in Fig. 3. Microvilli on enterocytes at the base of crypts, measuring about 0.7 pm in length, increase gradually and then rapidly as cells migrate out of crypts onto
I
I
I
0.6 crypt-villus
structure
junction
0.9
I
I
1.2
I
I 1.5
(mm)
Fig. 3. Microvillus development during enterocyte migration along chicken proximal small intestinal villi. Tissues taken from chickens which were either unselected (A), relaxed selected (0) or highly selected (0) for rapid growth were processed for electron microscopy and measurement of microvillus length as described in the text. Each point gives the mean of two determinations carried out on villi obtained from six chickens per group. The arrow shows the position of the crypt-villus junction. A single logistic growth curve has been used to fit all experimental findings.
61
Genetic effects on intestinal structure Table 2. Logistic constants
describing microvillus along chicken proximal
development during enterocyte small intestinal villi
migration
6 Chicken line
(1%
lhrl
Unselected
0.70 f 0.12
Relaxed
0.79 k 0.16
Highly
selected selected
0.79 * 0.09
0.016 k 0.005 [O.1691 0.018 f 0.009 [O.1821 0.014 f 0.003 [0.197]
60 & 20 117.91 56 f 35 [lO.S] 103 f 20 [lS.l]
I.55 kO.15 1.26+0.19 1.33+o.ll
Constants give calculated values + SEM for a, the initial microvillus length; c, subsequent microvillus growth; m, the position or time when elongation is half complete and 6, an exponential coefficient describing the distance or time dependency of microvillus elongation. Results in square brackets refer to time dependent values only.
the base of villi. Maximal microvillus length reached about 0.3 mm from the crypt-villus junction is maintained to the villus tip. No obvious difference is seen between results obtained from unselected, relaxed selected and highly selected birds. A common logistic growth curve fitting all experimental data gave calculated initial and final microvillus lengths of 0.74 and 2.12 pm with microvillus elongation becoming half complete 75 pm from the crypt-villus junction. Individual logistic growth constants calculated for each group of birds are given in Table 2. Calculated values for the initial length (a) and subsequent growth (c) of microvillus membranes are very similar for all groups of chickens. The position and time when microvillus elongation becomes half complete (m) and the coefficients describing this development (b) show greater variation. None of these differences are however statistically significant. Final comparison of residual errors obtained from fitting three instead of one curve to these results produced no significant increase in goodness of fit (F = 1.4). It is concluded from these findings that microvillus development does not respond to selection pressure for faster growth. Microvillus density on the surface of enterocytes as well as the length of individual microvilli will, of course, also affect the degree to which microvilli can modify villus surface area. Results obtained from measuring microvillus densities at various positions along the crypt-villus axis are given in Table 3. The number of microvilli measured per pm length of sectioned brush border plasma membrane also increased significantly between crypt base, cryptvillus junction and mid villus but there was no detectable difference in microvillus density between different groups of chickens. The microvillus diamTable 3. Microvillus
eter also remained constant under all conditions. Maximum amplification of villus surface area by microvilli was finally calculated by multiplying microvillus surface area by microvillus density measured in the mid villus region. The factor of 20 obtained from this calculation is equal to that reported previously for human intestine (Wilson, 1962). Dual control over microvilius development
Previously it has been shown that crypt depth (CD) and enterocyte migration rate (R) together determine the amount of microvillus elongation taking place during enterocyte development (Smith and Brown, 1989). Results showing the relationship existing between these three parameters, determined after addition of results obtained from the present three groups of chickens, are shown in Fig. 4. Combined results have been fitted by a straight line having a slope of 0.077 f 0.009 and an intercept of 0.004 + 0.001. These constants are not significantly different from those determined before inclusion of the present data (slope of 0.077 f 0.011; intercept of -0.005 f 0.001). These findings confirm the generality of the model put forward recently to define variables controlling enterocyte expression of microvillus structure. DISCUSSION
Adaptation of villus structure is normally thought to be mediated through changes in luminal nutrition, pancreatico-biliary secretions and hormonal factors (Dowling, 1982). Work with foetal isografts also shows an inherent ability of the small intestine to create and maintain regional differences in villus height (MacDonald and Ferguson, 1982). Any one or more of the above factors could be responsible for
density in chicken enterocyte
brush border
Microvillus Chicken line Unselected Relaxed
selected
Highly selected The number electron the villus length k positions diameter
Crypt base 3.9 [S: 3.8 [4: 3.4 [II:
f 0.3 - 1261 f 0.6 -2231 * 0.3 -2281
Crypt-villus 4.6 f [7: 4.0 * [3: 5.2 f [IO:
membranes
densities junction 0.2 O] 0.4 O] 0.2 01
Mid villus 5.3 + 0.2 [IO: 3661 5.7 + 0.3 [5: 6501 5.8 + 0.4 [9: 6501
of microvilli present in enterocyte brush border membranes was determined by microscopy near the crypt base, at the crypt-villus junction and at a point on where microvillus elongation was complete. Values give mean numbers per pm SEM. Values in brackets refer to numbers of chickens used and the mean of enterocytes measured in pm from the crypt-villus junction. Microvillus remained constant under all conditions (0.10 f 0.001 pm: 57 observations).
M. W. St+nrn et al.
62
01 0
Fig. 4. Combined effects of crypt size and enterocyte migration rate upon microvillus development. Reciprocal enterocyte migration rates (l/R) are plotted against microvillus elongation/crypt depth ratios (c/CD) for mammalian and chicken intestines. Addition of present results obtained from chickens which were unselected, relaxed selected and highly selected for rapid growth (0) have been added to a previously published data base (0; Smith and Brown, 1989) to produce a straight line relationship having a slope of 0.077 f 0.09 with an intercept of -0.004 + 0.001 (P < 0.001). creating the differences ported here for different
in villus surface area regroups of chickens. In spite
of this complexity there seems little doubt that increased presentation of nutrients to the small intestine constitutes one of the most potent ways to modify villus structure. Results supporting this statement come from collaborative work carried out between ourselves and others working on a variety of experimental models. One can for instance increase villus surface area by increasing food intake either directly or by hypothermia (Dauncey et al., 1983; Cremaschi et al., 1986); by feeding isocalorific diets containing different amounts of proteins (Syme, 1982; King et al., 1983) or by presenting similar amounts of nutrients to resected intestine (Menge et al., 1983; Chaves et al., 1987). Part at least of the differences noted in villus structure in the present work must have resulted from differences in food intake. Suspicion
that food intake alone cannot account for all of the noted changes arises from the parallel finding that all of the situations mentioned above, and several others referred to recently by Goodlad et al. (1987). are associated with increases in enterocyte migration or crypt cell production rates. There is in fact no situation to our knowledge where increased luminal nutrition can be shown to increase villus height without affecting crypt proliferation. In this case it is particularly interesting to note that both crypt size and enterocyte migration rate remain unchanged in unselected and relaxed selected birds at a time when villus surface areas and chicken growth rates increase two to threefold. These results strongly suggest that genetic selection, carried out empirically by commercial breeders, has succeeded in uncoupling
a well established link between crypt hyperplasia and villus structure. It is also interesting to note that further selection producing a 40% increase in growth rate between highly selected and relaxed selected chickens is accompanied by a 40% increase in villus area, crypt size and enterocyte migration rate (Table 1). This is entirely predictable from correlations already found to exist by Goodlad et al. (1987) between food intake, crypt cell production rate and intestinal weight. Genetic selection for appetite alone could account for these findings. Microvilli cause considerable amplification of villus surface area and microvillus length depends on crypt size and enterocyte migration rate (Smith et al., 1984; Smith and Peacock, 1989; Smith and Brown, 1989). Genetic selection for either or both of these parameters could affect the total surface area available for digestion and absorption of nutrients. There was, however, no evidence to suggest that microvillus numbers or developmental profiles for microvillus growth were changing between the three groups of birds. Crypt size increased by 40% in highly selected chickens and this would be predicted to increase microvillus length. This predicted increase appears to have been exactly matched in the present case by a 40% increase in enterocyte migration rate. The general equation providing a quantitative description for microvillus growth, M = 0.0016 CD + 0.077 CD/R, where M is the maximal
microvillus
length, CD is the
crypt depth and R is the enterocyte migration rate remained unchanged by addition of present results to previously published data (Fig. 4). This does not mean, however, that future selection could not produce chickens having larger crypts or lower rates of enterocyte migration. Indeed there is already some evidence to suggest that this might already have taken place in laying hens where the maximal microvillus length is 35% greater than that found in the present work. Finally it should be pointed out that the ability of enterocytes to absorb sugars and amino acids only occurs late during enterocyte migration to the tips of villi (Kinter and Wilson, 1965; Smith, 1985). Comparing absorptive function to total villus surface area in this case will be subject to error whenever there are adaptive changes in the ability of individual enterocytes to absorb amino acids. Preliminary findings suggest that villus tip enterocytes absorb alanine less efficiently in unselected compared with the other two types of chicken intestine (Mitchell and Smith, 1990). This could explain why differences in villus surface area between relaxed and unselected lines of bird still underestimate differences in growth rate (twofold difference for structure; threefold difference for growth rate). It is necessary to monitor developmental profiles for absorptive capacity along the villus, as well as obtain measurements of total villus area, in order to obtain a complete description of how genetic selection might affect intestinal function. Such studies are currently in progress. REFERENCES Chaves M., Smith M. W. and Williamson R. C. N. (1987) Increased activity of digestive enzymes in ileal enterocytes adapting to small bowel resection. GUI 28, 981-987.
Genetic effects on intestinal structure Cremaschi D., James P. S., Meyer G., Rossetti C. and Smith M. W. (1986) Intracellular potassium as a possible inducer of amino acid transport across hamster jejunal enterocytes. J. Physiol. (Lond.) 375, 107-119. Dauncey M. J., Ingram D. L., James P. S. and Smith M. W. (1983) Modification by diet and environmental temperature of enterocyte function in piglet intestine. J. Physiol. (Lond.) 341, 441452. Dowling R. H. (1982) Small bowel adaptation and its regulation. Stand. J. Gaslroenterol. 17, suppl. 74, 53-74. Ferraris R. P.. Lee P. P. and Diamond J. M. (1989) Origin of regional and species differences in intestinal’glucose uptake. Am. J. Physiol. 257, G689-G697. Goodlad R. A., Plumb J. A. and Wright N. A. (1987) The relationship between intestinal crypt cell production and intestinal water absorption measured in vitro in the rat. Clin. Sri. 72, 297-304.
James P. S., Smith M. W. and Tivey D. R. (1988) Singlevillus analysis of disaccharidase expression by different regions of the mouse intestine. J. Physiol. (Lond.) 401, 533-545.
King I. S., Paterson J. Y. F., Peacock M. A., Smith M. W. and Syme G. (1983). Effect of diet upon enterocyte differentiation in the rat jejunum. J. Physiol. (Lond.) 344, 46548
1.
Kinter W. B. and Wilson T. H. (1965) Autoradiographic study of sugar and amino acid absorption by everted sacs of hamster intestine. J. cell. Biol. 25, 19-39. MacDonald T. T. and Ferguson A. (1982) Regulation of villus height: the role of luminal factors in determining the villus height gradient of the mouse small intestine. In Mechanisms of Intestinal Adaptalion (Edited by Robinson
63
J. W. L., Dowling R. H. and Riecken E-O.), pp. 47-53. MTP Press, Lancaster. Menge H., Sepulveda F. V. amd Smith M. W. (1983) Cellular adaptation of amino acid transport following intestinal resection in the rat. J. Physiol. (Lond.) 334, 213-223.
Mitchell M. A. and Smith M. W. (1990) Jejunal alanine uptake and structural adaptation in response to genetic selection for growth rate in the domestic fowl (Gal/us domesticus).
J. Physiol.
(Lond.) 424.
Smith M. W. (1985) Expression of digestive and absorptive function in differentiating enteroctyes. Ann. Rev. Physiol. 47, 247-260.
Smith M. W. and Brown D. (1989) Dual control over microvillus elongation during enterocyte development. Comp. Biochem.
Physiol. 93A, 623-628.
Smith M. W., Paterson J. Y. F. and Peacock M. A. (1984) A comprehensive description of brush border membrane development applying to enterocytes taken from a wide variety of mammalian species. Comp. Biochem. Physiol. VA, 655462.
Smith M. W. and Peacock M. A. (1989) Comparative aspects of microvillus development in avian and mammalian enterocytes. Comp. Biochem. Physiol. 93A, 617422. Syme G. (1982) The effect of protein-deficient isoenergetic diets on the growth of rat jejunal mucosa. Br. J. Nutr. 48, 25-36.
Tivey D. R. and Smith M. W. (1989) Cytochemical analysis of single villus peptidase activities in pig intestine during neonatal development. Hisfochem. J. 21, 601-608. Wilson T. H. (1962) In Intestinal Absorption (Edited by Wilson T. H.), pp. I-19. W. B. Saunders, Philadelphia.
1990
0300-9629/90 $3.00 + 0.00 0 1990 Pergamon Press pie
EFFECTS OF GENETIC SELECTION ON GROWTH RATE AND INTESTINAL STRUCTURE IN THE DOMESTIC FOWL (GALLUS DOMESTICU,!2) M.W.
SMITH,*M. A. MITCHELL?and M. A. PEACOCK* *AFRC Institute of Animal Physiology and Genetics Research, Babraham, Cambridge CB2 4AT, UK. Telephone: 0223-832312;and tRosiin, Midlo~ian EH25 9PS, UK. Telephone: 031-440-2726
Abstract-I.
Pieces of small intestine taken from chickens subjected previously to continuous selection, relaxed selection or no selection for rapid growth were used to estimate villus surface area and microvillus development to determine what effects genetic selection might have on factors controlling intestinal function. 2. Crypt size and the rates at which enterocytes migrated out of crypts were also measured, after injection of tritiated thymidine, to determine the time course of microvillus elongation. 3. Differences in growth rates measured between highly selected, relaxed selected or unselected birds were found to be correlated with parallel changes in villus surface area. Selection for growth did not change the density, dimensions or pattern of development of enterocyte microviili. Mi~rovilli did, however, produce a maximal 20-fold increase in villus surface area under all conditions. 4. Crypt size and enterocyte migration rates did not vary significantly between tissue taken from unselected and relaxed selected chickens. Tissue taken from highly selected birds had a crypt size and enterocyte migration rate 40% higher than values found for the other two groups of chickens. 5. The possibility that early genetic selection increased growth potential by uncoupling diet-induced changes on crypt hyperplasia from secondary effects on villus structure, and that later selection increased growth potential by increasing appetite, is discussed.
MATERIALS AND METHODS
Changes in enterocyte function and villus surface area occurring during the lifetime of the animal or, more permanently, through chance mutation favouring survival, together account for the ability of the small intestine to modify nutrient capture and energy metabolism (James et al., 1988; Tivey and Smith, 1989). Microvilli present on the enterocyte surface further amplify the area of villus available for the digestion and absorption of nutrients (Wilson, 1962). Changes at this level have also been found to be linearly related to the metabolic live mass of different mammalian species (Ferraris et al., 1989). The present situation is therefore one where interest in determining both villus and microvillus structure is increasing in studies designed to identify factors controlljng intestinal function. Neglected so far in this type of consideration is work involving chickens, a species genetically selected by commercial breeders in order to improve growth performance. Present work uses chickens subjected previously to different levels of genetic selection to test the hypothesis that accelerated weight gain might be associated with major changes in intestinal function. Results obtained showed vilius area, but not microvillus structure, to be directly related to chicken growth rates. The possibility that increases in villus area enable faster growing birds to sustain increased demands for nutrient assimilation is discussed.
Four-week-old female chickens used in the present experiments came from three lines exposed to different degrees of selection pressure for rapid growth rate and feed conversion efficiency. A modem commercial line was employed which has been continuously selected upon these criteria until the present day (designated highly selected). A second line had been selected for growth rate until 1972 and then selection was discontinued (relaxed selected line). A third brownleghorn line had never been selected for increased growth rate (J-line unselected). Newly hatched chickens from all three sources were originally maintained in brooders kept at an environmental temperature of 30°C. This temperature was gradually reduced to 22°C over a 2 week period after which birds were transferred to individual cages in climatic chambers, where they were kept under controlled conditions of tem~rature and fighting (2210S”C: 14 hr Ii~t~lO hr dark). -A commercial- broie; m&h was fed ad ~~~it~~ thoughout to chickens allowed free access to water. Weight gain in these animals was monitored by daily weighing during the week prior to the experiment.
Small intestines removed from chickens killed by cervical dislocation were measured and samples of tissue taken 25, 50 and 75% along the tract for further analysis. Initial preparation involved flushing ail tissue samples with phosphate buffered saline to remove intestinal contents before dividing each portion into three for later dete~ination of villus structure by image analysis, enterocyte migration rate by autoradiography and microvillus development by electron microscopy. The depth of individual crypts was also
M. W. SMITH el al.
58
measured by image analysis of haematoxylin and eosin stained sections of material prepared for autoradiography. The amount by which enterocyte microvilli increased villus surface area was assessed separately by measuring the diameter, height and number of microvilli present per standard length of tissue section by electron microscopy. Complete analysis of the above parameters was restricted to the most proximal region of the small intestine. Estimates of enterocyte migration rate, villus and crypt structure, were carried out at all three positions along the gut. Measurement
of villus surface area
Intact villi microdissected from tissue fixed previously in phosphate buffer containing 4% v/v glutaraldehyde and 2% w/v sucrose, using fine Monoject hypodermic needles (0.5 x 16 mm), were mounted on microscope slides in a 1: 1 mixture of glycerol and phosphate-buffered saline. These villi were then flattened by applying a I kg force to a 2 x 2 cm cover-slip and the edges of the cover-slip sealed with Tipp-Ex to preserve villus conformation and prevent dehydration. The height and surface area of these villi was then determined by transferring images through a TV camera attached to a Leitz microscope into a Magiscan 2A image analyser. Enhanced contrast of these images was provided by passing light from the microscope lamp through a Glen Spectra narrow band filter. Values of surface area obtained using the standard Magiscan MENU program were finally doubled, it being assumed that each villus consisted of two flat sheets bent to meet each other only in the region of the villus perimeter. Measurement
of enterocyte migration rate
Initial labelling of crypt cells and subsequent migration of labelled enterocytes onto villi was determined by injecting tritiated thymidine intravenously (37 kBq/g body wt) into 42 chickens (14 per group). Injected animals were then killed 2, 24, 48 or 72 hr later to obtain tissue samples for fixation in sucrose-glutaraldehyde buffer. Fixed tissue washed in phosphate buffered saline was then dehydrated and embedded in glycol-methacrylate for sectioning on a Reichert-Jung Autocut microtome. Cut sections were then coated with Ilford K2 photographic emulsion for later development and enhancement of Ag grain densities. Both exposure and enhancement times needed to prepare suitable autoradiographs were much longer than reported previously when injecting tritiated thymidine intraperitoneally (Smith and Brown, 1989). This latter route of injection, which probably gives a long tissue exposure to label, is to be preferred to the intravenous route of administration. Prepared autoradiographs were later stained with eosin and the positions of the leading edges of labelled cells determined by image analysis. The distance migrated by labelled cells was linearly related to the time after thymidine injection. Regression analysis of this data gave calculated migration rates in pmm/hr for the three groups of chickens. Elecrron microscopical measurement of microvillus slructure Samples of small intestine taken from the proximal region of unselected, relaxed selected and highly selected chickens were cut open, pinned out onto cork and fixed flat in phosphate buffer containing 4% v/v glutaraldehyde and 2% w/v sucrose. Fixed tissues were then cut into small strips for post-fixation in Verona1 buffer containing 1% w/v osmium tetroxide. Post-fixed samples were finally dehydrated, embedded in Araldite and sectioned for examination in a Philips 400 electron microscope. Only those areas of tissue showing appropriately sectioned microvilli were used for measurement of microvillus length and diameter. Individual positions on the crypt-villus axis chosen for measurement of microvillus structure were also recorded using the cryptvillus junction as a reference point. Microvillus densities in crypts, at the crypt-villus junction and at positions on villi where microvillus elongation was
complete were measured separately by counting the number of microvilli present in a standard length of sectioned tissue. These values were then used together with estimates of microvillus length and diameter to calculate microvillus surface area per villus. Logistic curve analysis of microvilhs elongation The method adopted for fitting logistic growth curves to obtain constants describing quantitative aspects of microvillus formation during enterocyte migration from crypt base to villus tip was identical to that described previously (Smith et al., 1984; Smith and Peacock, 1989). Briefly it consisted of fitting logistic growth curves having the form y = a + c/{ I + exp[ - b(x - m)]} to measurements of microvillus length obtained along the whole crypt-villus axis of unselected, relaxed selected and highly selected chickens and comparing these results with a single logistic curve obtained from fitting combined results derived from all chickens. In this formula, a is the calculated initial length of the microvillus, c is the microvillus growth that occurs subsequently, m is the position or time when each enterocyte half completes this growth phase and b is an exponential coefficient. Slaristical analysis Unpaired Student’s r-tests were used to assess the statistical significance of differences found between mean results f SEM. Standard errors associated with calculated logistic curve constants refer to the goodness of fit of curves to the experimental data rather than differences existing between individual chickens. The F test was used to assess whe*tier the error involved in fitting three separate growth curves to data obtained from the three groups of chickens was significantly different from that found when fitting a single growth curve to pooled data. Materials The [mefhyl-3 H] thymidine (I .5-2.2 TBq/mmol) came from Amersham International plc, Amersham, UK. All other reagents used were of AR Grade. RESULTS
Genetic selection for chicken growth rate Chickens classified as being unselected, relaxed selected or highly selected for rapid growth were found to weigh 186 f 5, 523 f 10 and 775 f 20 g 3 weeks after hatching (means f SEM; 21-23 birds per group). These weights had increased significantly to 249 rf: 6, 711 rf: 12 and 1048 + 29g 1 week later. Weight gains measured daily in chickens during the period 34 weeks after hatching were related linearly to age of chicken in all three groups (Fig. 1). Fitting regression lines to these data gave daily weight gains of 10.6 f 0.4, 30.1 + 1.7 and 43.3 f 2.2 g for unselected, relaxed selected and highly selected birds respectively. It is concluded from these initial findings that all three groups of chickens are growing normally at this stage in development and that differences in absolute growth rates are sufficiently great as to warrant further investigation.
Genetic selection for intestinal structure and enterocyte replacement
Changes in intestinal structure were tested for in the three lines of chickens chosen for study by measuring villus surface area, villus height, crypt depth and enterocyte migration rate as described in the text. Results obtained for villus surface area are summarized in Fig. 2.
Genetic effects on intestinal structure
59
0.3 r
I
22
I
I
I
I
I
23
24
25
25
2l
I
25
Fig. 1. Effect of genetic selection upon chicken growth rate. Chickens which were either unselected (A), relaxed selected (0) or highly selected (0) for rapid growth were allowed free access to food and water until 4 weeks of age when they were killed for experiment. Values give mean weight gains measured during the week prior to experiment (21-23 chickens per group). Villus surface area from unselected chickens was less than that determined in relaxed selected birds. Largest villus surface area occurted in chickens
highly selected for growth. These changes took place without altering the overall shape of villi which remained roughly rectangular throughout. This constancy of proportionality allowed one to fit a single curve to all experimental findings relating surface area (A: mm2) to villus height (H: mm) using the
A = 1.26H’.‘r. Values for mean surface area determined for the three groups of birds, together with other measurements of intestinal structure, are summarized in Table 1. Villi taken from birds where selection had been relaxed had twice the surface area of those determined in unselected birds (1.6 vs 0.8 mm2). This increase took place without change in crypt depth or enterocyte migration rate. A further 50% increase in
equation
4-
3-
A-
d 4
91
a
s
2-
3 >
l-
1.0 Villus
length
1.5
PO
(mm)
Fig. 2. ElTect of genetic selection upon the height and surface area of chicken intestinal villi. Microdissected villi from the proximal small intestine were analysed from chickens which were unselected (A), relaxed selected (0) or highly selected (0) for rapid growth. The fitted curve relating surface area (A) to villus height (If) has the form A = 1.269H’,“.
M. W. St+ttr~ et al.
60 Table I. Effect of genetic selection
Relaxed selected Highly
selected
and enterocyte
kinetics
in chicken
VillUS height (mm)
Villus area (mm’)
Crypt depth (mm)
Enterocyte migration rate (rmlhr)
0.85 f 0.04 (15:45) I .09 * 0.05 (15:45) I .3l k 0.06 (15:45)
0.78 ? 0.03 (15:45) I.61 k 0.08 (15:45) 2.27 + 0.18 (15:45)
0.20 &-0.01 (15:150) 0.19~0.01 (15:150) 0.27 f 0.01 (15:257)
IO.5 * 1.0 (14:282) 10.1 _+0.6 (14:194) 14.0 It 0.6 (13:257)
Chicken line Unselected
upon intestinal structure proximal small intestine
Measurements of villus height and area were obtained by image analysis of microdissected tissue taken from chickens injected previously with tritiated thymidine. Enterocyte migration rate and crypt depth measurements were carried out on eosin stained autoradiographs prepared from adjacent pieces of tissue as described in the text. Values give means + SEM. Numbers in brackets show the number of chickens used and villi analysed respectively.
villus surface was seen in birds highly selected for rapid growth (2.3 vs 1.6 mm2). This change was associated with significant increases in both crypt depth and enterocyte migration rate (P < 0.001 in both cases). Further morphometic and enterocyte kinetic measurements, carried out in mid-distal regions of the small intestine, showed enterocyte migration to fall to 40-50% of that measured in the proximal small intestine (5.4 vs 10.5; 6.9 vs 10.1 and 10.1 vs 14.0 pm/ hr for unselected, relaxed selected and highly selected birds). These changes were not associated with any obvious alteration in crypt depth (0.18 vs 0.20; 0.19 vs 0.19 and 0.24 vs 0.27 mm, respectively). It is concluded that selective line-dependent changes in intestinal structure and enterocyte migration do take place along the whole length of the small intestine.
I - 0.3
I
I
I
0
I
I
0.3 Distance
from
Changes in villus surface area are moreover roughly proportional to differences noted earlier in chicken growth rates (Table 1; Fig. 1). Genetic selection for microvillus
Microvilli are known to cause large increases in apparent villus surface area in mammalian small intestines (Wilson, 1962; Ferraris et al., 1989). It was therefore considered necessary to measure development of microvillus structure in these three types of bird to test whether selection pressure could also be detected at this level of analysis. Results obtained from carrying out these measurements are summarized in Fig. 3. Microvilli on enterocytes at the base of crypts, measuring about 0.7 pm in length, increase gradually and then rapidly as cells migrate out of crypts onto
I
I
I
0.6 crypt-villus
structure
junction
0.9
I
I
1.2
I
I 1.5
(mm)
Fig. 3. Microvillus development during enterocyte migration along chicken proximal small intestinal villi. Tissues taken from chickens which were either unselected (A), relaxed selected (0) or highly selected (0) for rapid growth were processed for electron microscopy and measurement of microvillus length as described in the text. Each point gives the mean of two determinations carried out on villi obtained from six chickens per group. The arrow shows the position of the crypt-villus junction. A single logistic growth curve has been used to fit all experimental findings.
61
Genetic effects on intestinal structure Table 2. Logistic constants
describing microvillus along chicken proximal
development during enterocyte small intestinal villi
migration
6 Chicken line
(1%
lhrl
Unselected
0.70 f 0.12
Relaxed
0.79 k 0.16
Highly
selected selected
0.79 * 0.09
0.016 k 0.005 [O.1691 0.018 f 0.009 [O.1821 0.014 f 0.003 [0.197]
60 & 20 117.91 56 f 35 [lO.S] 103 f 20 [lS.l]
I.55 kO.15 1.26+0.19 1.33+o.ll
Constants give calculated values + SEM for a, the initial microvillus length; c, subsequent microvillus growth; m, the position or time when elongation is half complete and 6, an exponential coefficient describing the distance or time dependency of microvillus elongation. Results in square brackets refer to time dependent values only.
the base of villi. Maximal microvillus length reached about 0.3 mm from the crypt-villus junction is maintained to the villus tip. No obvious difference is seen between results obtained from unselected, relaxed selected and highly selected birds. A common logistic growth curve fitting all experimental data gave calculated initial and final microvillus lengths of 0.74 and 2.12 pm with microvillus elongation becoming half complete 75 pm from the crypt-villus junction. Individual logistic growth constants calculated for each group of birds are given in Table 2. Calculated values for the initial length (a) and subsequent growth (c) of microvillus membranes are very similar for all groups of chickens. The position and time when microvillus elongation becomes half complete (m) and the coefficients describing this development (b) show greater variation. None of these differences are however statistically significant. Final comparison of residual errors obtained from fitting three instead of one curve to these results produced no significant increase in goodness of fit (F = 1.4). It is concluded from these findings that microvillus development does not respond to selection pressure for faster growth. Microvillus density on the surface of enterocytes as well as the length of individual microvilli will, of course, also affect the degree to which microvilli can modify villus surface area. Results obtained from measuring microvillus densities at various positions along the crypt-villus axis are given in Table 3. The number of microvilli measured per pm length of sectioned brush border plasma membrane also increased significantly between crypt base, cryptvillus junction and mid villus but there was no detectable difference in microvillus density between different groups of chickens. The microvillus diamTable 3. Microvillus
eter also remained constant under all conditions. Maximum amplification of villus surface area by microvilli was finally calculated by multiplying microvillus surface area by microvillus density measured in the mid villus region. The factor of 20 obtained from this calculation is equal to that reported previously for human intestine (Wilson, 1962). Dual control over microvilius development
Previously it has been shown that crypt depth (CD) and enterocyte migration rate (R) together determine the amount of microvillus elongation taking place during enterocyte development (Smith and Brown, 1989). Results showing the relationship existing between these three parameters, determined after addition of results obtained from the present three groups of chickens, are shown in Fig. 4. Combined results have been fitted by a straight line having a slope of 0.077 f 0.009 and an intercept of 0.004 + 0.001. These constants are not significantly different from those determined before inclusion of the present data (slope of 0.077 f 0.011; intercept of -0.005 f 0.001). These findings confirm the generality of the model put forward recently to define variables controlling enterocyte expression of microvillus structure. DISCUSSION
Adaptation of villus structure is normally thought to be mediated through changes in luminal nutrition, pancreatico-biliary secretions and hormonal factors (Dowling, 1982). Work with foetal isografts also shows an inherent ability of the small intestine to create and maintain regional differences in villus height (MacDonald and Ferguson, 1982). Any one or more of the above factors could be responsible for
density in chicken enterocyte
brush border
Microvillus Chicken line Unselected Relaxed
selected
Highly selected The number electron the villus length k positions diameter
Crypt base 3.9 [S: 3.8 [4: 3.4 [II:
f 0.3 - 1261 f 0.6 -2231 * 0.3 -2281
Crypt-villus 4.6 f [7: 4.0 * [3: 5.2 f [IO:
membranes
densities junction 0.2 O] 0.4 O] 0.2 01
Mid villus 5.3 + 0.2 [IO: 3661 5.7 + 0.3 [5: 6501 5.8 + 0.4 [9: 6501
of microvilli present in enterocyte brush border membranes was determined by microscopy near the crypt base, at the crypt-villus junction and at a point on where microvillus elongation was complete. Values give mean numbers per pm SEM. Values in brackets refer to numbers of chickens used and the mean of enterocytes measured in pm from the crypt-villus junction. Microvillus remained constant under all conditions (0.10 f 0.001 pm: 57 observations).
M. W. St+nrn et al.
62
01 0
Fig. 4. Combined effects of crypt size and enterocyte migration rate upon microvillus development. Reciprocal enterocyte migration rates (l/R) are plotted against microvillus elongation/crypt depth ratios (c/CD) for mammalian and chicken intestines. Addition of present results obtained from chickens which were unselected, relaxed selected and highly selected for rapid growth (0) have been added to a previously published data base (0; Smith and Brown, 1989) to produce a straight line relationship having a slope of 0.077 f 0.09 with an intercept of -0.004 + 0.001 (P < 0.001). creating the differences ported here for different
in villus surface area regroups of chickens. In spite
of this complexity there seems little doubt that increased presentation of nutrients to the small intestine constitutes one of the most potent ways to modify villus structure. Results supporting this statement come from collaborative work carried out between ourselves and others working on a variety of experimental models. One can for instance increase villus surface area by increasing food intake either directly or by hypothermia (Dauncey et al., 1983; Cremaschi et al., 1986); by feeding isocalorific diets containing different amounts of proteins (Syme, 1982; King et al., 1983) or by presenting similar amounts of nutrients to resected intestine (Menge et al., 1983; Chaves et al., 1987). Part at least of the differences noted in villus structure in the present work must have resulted from differences in food intake. Suspicion
that food intake alone cannot account for all of the noted changes arises from the parallel finding that all of the situations mentioned above, and several others referred to recently by Goodlad et al. (1987). are associated with increases in enterocyte migration or crypt cell production rates. There is in fact no situation to our knowledge where increased luminal nutrition can be shown to increase villus height without affecting crypt proliferation. In this case it is particularly interesting to note that both crypt size and enterocyte migration rate remain unchanged in unselected and relaxed selected birds at a time when villus surface areas and chicken growth rates increase two to threefold. These results strongly suggest that genetic selection, carried out empirically by commercial breeders, has succeeded in uncoupling
a well established link between crypt hyperplasia and villus structure. It is also interesting to note that further selection producing a 40% increase in growth rate between highly selected and relaxed selected chickens is accompanied by a 40% increase in villus area, crypt size and enterocyte migration rate (Table 1). This is entirely predictable from correlations already found to exist by Goodlad et al. (1987) between food intake, crypt cell production rate and intestinal weight. Genetic selection for appetite alone could account for these findings. Microvilli cause considerable amplification of villus surface area and microvillus length depends on crypt size and enterocyte migration rate (Smith et al., 1984; Smith and Peacock, 1989; Smith and Brown, 1989). Genetic selection for either or both of these parameters could affect the total surface area available for digestion and absorption of nutrients. There was, however, no evidence to suggest that microvillus numbers or developmental profiles for microvillus growth were changing between the three groups of birds. Crypt size increased by 40% in highly selected chickens and this would be predicted to increase microvillus length. This predicted increase appears to have been exactly matched in the present case by a 40% increase in enterocyte migration rate. The general equation providing a quantitative description for microvillus growth, M = 0.0016 CD + 0.077 CD/R, where M is the maximal
microvillus
length, CD is the
crypt depth and R is the enterocyte migration rate remained unchanged by addition of present results to previously published data (Fig. 4). This does not mean, however, that future selection could not produce chickens having larger crypts or lower rates of enterocyte migration. Indeed there is already some evidence to suggest that this might already have taken place in laying hens where the maximal microvillus length is 35% greater than that found in the present work. Finally it should be pointed out that the ability of enterocytes to absorb sugars and amino acids only occurs late during enterocyte migration to the tips of villi (Kinter and Wilson, 1965; Smith, 1985). Comparing absorptive function to total villus surface area in this case will be subject to error whenever there are adaptive changes in the ability of individual enterocytes to absorb amino acids. Preliminary findings suggest that villus tip enterocytes absorb alanine less efficiently in unselected compared with the other two types of chicken intestine (Mitchell and Smith, 1990). This could explain why differences in villus surface area between relaxed and unselected lines of bird still underestimate differences in growth rate (twofold difference for structure; threefold difference for growth rate). It is necessary to monitor developmental profiles for absorptive capacity along the villus, as well as obtain measurements of total villus area, in order to obtain a complete description of how genetic selection might affect intestinal function. Such studies are currently in progress. REFERENCES Chaves M., Smith M. W. and Williamson R. C. N. (1987) Increased activity of digestive enzymes in ileal enterocytes adapting to small bowel resection. GUI 28, 981-987.
Genetic effects on intestinal structure Cremaschi D., James P. S., Meyer G., Rossetti C. and Smith M. W. (1986) Intracellular potassium as a possible inducer of amino acid transport across hamster jejunal enterocytes. J. Physiol. (Lond.) 375, 107-119. Dauncey M. J., Ingram D. L., James P. S. and Smith M. W. (1983) Modification by diet and environmental temperature of enterocyte function in piglet intestine. J. Physiol. (Lond.) 341, 441452. Dowling R. H. (1982) Small bowel adaptation and its regulation. Stand. J. Gaslroenterol. 17, suppl. 74, 53-74. Ferraris R. P.. Lee P. P. and Diamond J. M. (1989) Origin of regional and species differences in intestinal’glucose uptake. Am. J. Physiol. 257, G689-G697. Goodlad R. A., Plumb J. A. and Wright N. A. (1987) The relationship between intestinal crypt cell production and intestinal water absorption measured in vitro in the rat. Clin. Sri. 72, 297-304.
James P. S., Smith M. W. and Tivey D. R. (1988) Singlevillus analysis of disaccharidase expression by different regions of the mouse intestine. J. Physiol. (Lond.) 401, 533-545.
King I. S., Paterson J. Y. F., Peacock M. A., Smith M. W. and Syme G. (1983). Effect of diet upon enterocyte differentiation in the rat jejunum. J. Physiol. (Lond.) 344, 46548
1.
Kinter W. B. and Wilson T. H. (1965) Autoradiographic study of sugar and amino acid absorption by everted sacs of hamster intestine. J. cell. Biol. 25, 19-39. MacDonald T. T. and Ferguson A. (1982) Regulation of villus height: the role of luminal factors in determining the villus height gradient of the mouse small intestine. In Mechanisms of Intestinal Adaptalion (Edited by Robinson
63
J. W. L., Dowling R. H. and Riecken E-O.), pp. 47-53. MTP Press, Lancaster. Menge H., Sepulveda F. V. amd Smith M. W. (1983) Cellular adaptation of amino acid transport following intestinal resection in the rat. J. Physiol. (Lond.) 334, 213-223.
Mitchell M. A. and Smith M. W. (1990) Jejunal alanine uptake and structural adaptation in response to genetic selection for growth rate in the domestic fowl (Gal/us domesticus).
J. Physiol.
(Lond.) 424.
Smith M. W. (1985) Expression of digestive and absorptive function in differentiating enteroctyes. Ann. Rev. Physiol. 47, 247-260.
Smith M. W. and Brown D. (1989) Dual control over microvillus elongation during enterocyte development. Comp. Biochem.
Physiol. 93A, 623-628.
Smith M. W., Paterson J. Y. F. and Peacock M. A. (1984) A comprehensive description of brush border membrane development applying to enterocytes taken from a wide variety of mammalian species. Comp. Biochem. Physiol. VA, 655462.
Smith M. W. and Peacock M. A. (1989) Comparative aspects of microvillus development in avian and mammalian enterocytes. Comp. Biochem. Physiol. 93A, 617422. Syme G. (1982) The effect of protein-deficient isoenergetic diets on the growth of rat jejunal mucosa. Br. J. Nutr. 48, 25-36.
Tivey D. R. and Smith M. W. (1989) Cytochemical analysis of single villus peptidase activities in pig intestine during neonatal development. Hisfochem. J. 21, 601-608. Wilson T. H. (1962) In Intestinal Absorption (Edited by Wilson T. H.), pp. I-19. W. B. Saunders, Philadelphia.
