Main

The epithelial cells lining the colonic lumen perform both a crucial barrier function and are responsible for salt and water reabsorption. The fluid content of the colonic lumen is maintained by a balance of ion-absorptive and ion-secretory mechanisms. These transport mechanisms are regulated by hormones and neuromodulators(1–3) and undergo ontogenic changes(4–8). Abnormalities in epithelial ion transport in young and adult animals can result in disorders ranging from diarrhea(9) to cystic fibrosis(10). In addition, aberrations in cell to cell and cell to matrix attachment may be the central defect in inflammatory bowel disease in humans(9). Traditionally the study of colonic epithelial function has relied on the use of mucosal epithelial sheets in Ussing chambers, or of colonic cell lines such as T-84, Caco-2, and HT-29 cells. Many of the available cell lines have been derived from colonic adenocarcinomas(11). These approaches are limited by the fact that epithelial sheet preparations are comprised of epithelial and subepithelial elements and transformed cell lines may not accurately reflect native colonocyte function. There has been some success in establishing permanent epithelial cell lines from the small intestine of suckling animals(12–14) and from canine colon(15). Recently, there is some preliminary information on the development of nontransformed cell lines from the human colon(16).

A model system that is comprised of nontransformed cells and is devoid of subepithelial elements is that of a primary culture of isolated colonocytes. However, the procurement of isolated colonocytes and their maintenancein vitro has proven to be somewhat difficult. Studies on the small intestine indicate that enterocytes of young, suckling animals lend themselves to culture more readily than those of adults(12). Although we(17) and other investigators have placed adult mammalian colonocytes into primary culture with some success(15, 18–21), there have been no reports on the isolation and primary culture of newborn mammalian colonocytes. Consequently there is a need to develop an in vitro model so that developmental changes in the colon can be studied at the cellular level. An important component of developing a viable and functional culture of colonocytes is that their interaction with the subcellular matrix is critical for mucosal architecture and cellular differentiation(22, 23). Colonocytes attach poorly to plastic and glass but benefit significantly from the presence of attachment proteins, such as laminin and fibronectin(24). It is therefore necessary to define the matrix that optimizes colonocyte attachment. Independent of long-term culture considerations, colonocyte attachment studies may also be of value in the investigation of diseases, such as inflammatory bowel disease, in which abnormal cell-to-matrix interactions exist.

We used the rabbit colon as a model to compare the growth and attachment with various extracellular matrices of colonocytes obtained from different developmental ages. Colonocytes were obtained from suckling (7-9 d), weanling(25-28 d), and adult (3-4 mo) animals. First, we defined an optimal enzyme mixture (0.05% Pronase, 0.015% collagenase IV, 0.023% DTT for 60 min) for the isolation of newborn and weanling rabbit colonocytes. In parallel we confirmed our earlier findings(17) that the optimal conditions for isolation of adult colonocytes are 0.1% Pronase, 0.03% collagenase IV, and 0.07% DTT for 90 min. The viability and cell yield/g of mucosa varied with age. Second, we evaluated growth and proliferation of these colonocytes in primary cultures over a period of 72 h and observed development stage-related differences. Third, we systematically evaluated different matrices for attachment and demonstrated that, in newborn and weanling rabbits, colonocytes showed a high percentage of attachment to both Matrigel1 (Collaborative Research, Bedford, MA) and collagen IV. However, the overall tendency of colonocytes to attach to matrices showed a gradual decline as the age increases.

METHODS

Animals. New Zealand White newborn rabbits (7-9 d), weanlings(25-28 d), and adult male rabbits (2-3 kg) were procured from Lesser Rabbits, Delfield, WI. The newborn rabbits were housed with their mothers, and all animals were maintained on rabbit chow and water ad libitum. All animals were housed according to regulations of the Institutional Animal Care Committee and the guidelines of the National Institutes of Health. The animals were killed by phenobarbital overdose, and the distal colon was excised from the anal verge to the splenic flexure.

Cell isolation. The colon was opened along the mesenteric border and rinsed thoroughly in ice-cold (4°C) lactated Ringer's solution(Baxter Healthcare Corporation, Valencia, CA) (LRG) and antibiotics. The mucosa was blotted with tissue to remove mucus, and the mucosal layer was then isolated by blunt dissection. The mucosal layer was rinsed in cold LRG and then minced with crossed scalpels into small pieces. The colonic mucosa (of one adult rabbit, two weanlings, or four newborns) was placed in a container with 40 mL of enzyme digestion solution. We had previously demonstrated(17) that a digestion solution, containing 0.1% Pronase, 0.03% collagenase IV, and 0.07% DTT, was optimal for the isolation of adult colonocytes. In the present study we compared the efficacy of four different concentrations of the above components for the isolation of newborn, weanling, and adult colonocytes. The compositions of solutions I-IV are listed inTable 1. All solutions contained antibiotics (25 mg/L ampicillin, 125 mg/L penicillin, 270 mg/L streptomycin, and 1.25 mg/L amphotericin).

Table 1 Effect of different digestion solutions on colonocyte yield and viability
Full size table

The mucosa in enzyme solution was subjected to brisk agitation (175 cycles/min) in a 37°C gyrorotary orbit shaker bath (Labline Instruments Inc., Melrose Park, IL) for 90 min for adult colons and 60 min for colons of newborn animals and weanlings. Preliminary studies had demonstrated that these were the optimal time periods for the respective groups (data not shown). After agitation, the solution was strained twice through sterile gauze and then centrifuged at 4000 × g for 5 min. The pellet was washed once by resuspending in 20 mL of LRG and recentrifuging at 4000 ×g for 5 min. To enrich for crypts, as described by Benya et al.(17), the pellet was washed in LRG and centrifuged twice at 400 × g for 15 min. After the second centrifugation, the cells were resuspended in 1 mL of LRG at 4°C and placed in primary culture at 37°C.

Primary culture. Colonocytes were cultured in Ham's F-12 medium supplemented with 20% fetal bovine serum, 86.1 pmol/L insulin, 4 mmol/L L-glutamine, 1.1 μmol/L hydrocortisone, 1 mmol/L sodium butyrate, and antibiotics at the concentrations used in the digestion solutions described above. Colonocytes were preplated on plastic for 30 min to decrease contaminating fibroblasts, and then 500 μL of cell suspension containing 5× 105 cells were plated per each well (8 mm) in 24-well plates. Unless otherwise indicated, cells were plated in triplicate for each variable. The plates were placed in a 6% CO2 incubator at 37°C for different time periods up to 72 h.

Assessment of viability, growth, and proliferation. To assess viability in the preparations containing colonic epithelial crypts or clusters, aliquots of cells were triturated to disaggregate the clumps and thereby facilitate counting. Cell viability was assessed by monitoring cytoplasmic exclusion of 0.1% trypan blue(25). Viability was assessed in all age groups at 0, 24, 48, and 72 h. We had previously demonstrated that there is no difference in the cell viability assessed by trypan blue exclusion or by other viability markers such as fluorescein diacetate(17).

Growth and proliferation were assessed by cell count and protein and DNA assay at 0, 24, 48, and 72 h postplating. Cells were plated on 24-well plastic plates, or on Matrigel- or collagen IV-coated cell culture inserts (see next section). Cultured cells were harvested by first aspirating the unattached cells and reserving them on ice. The wells were then exposed to 250 μL of 0.1% trypsin for 15 min to release the attached cells. Trypsinization was needed to get an accurate quantitation of attached cells. Attached cells and cells in suspension were counted separately using a hemocytometer. Protein was assayed according to the method of Lowry et al.(26) and DNA by a fluorometric method described by Labarca and Paigen(27).

Cell attachment. Commercially available BIOCOAT (Collaborative Research) cell culture inserts precoated with different extracellular matrices were used. The following matrices were evaluated and compared with plastic: collagen type I, collagen type IV, Matrigel, laminin, and fibronectin. The inserts were rehydrated by addition of 0.5 mL of culture medium inside the inserts. After 30 min at room temperature, the medium was aspirated. Six hundred microliters of medium were added to each well in which the inserts sit. Five hundred microliters of cell suspension in medium were added to each insert.

To determine the best matrix for attachment, 5 × 105 cells were plated in duplicate on different matrices and placed in a 6% CO2 incubator at 37°C for 24 h. The media were aspirated and the nonattached cells counted. The cells in inserts/wells were washed with PBS three times and then exposed to 0.1% trypsin (250 μL) for 15 min. The trypsin was aspirated, and the detached cells were counted in a hemocytometer.

The time course of attachment was examined in detail, using the best matrix. Colonocytes were plated in quadruplicate on Matrigel and harvested using 0.1% trypsin at 2, 4, 16, 24, 48, and 72 h, and the attached and nonattached cells were harvested and counted as described above.

Immunocytochemistry and histology. Cultured colonocytes were fixed and stained for intermediate filaments according to the method of Yanget al.(28). Briefly, colonocytes grown on glass coverslips coated with collagen IV were washed in PBS with 0.2% BSA and then fixed in ice-cold methanol. After air drying, the cells were incubated with goat serum, washed, and exposed to mouse polycolonal anti-cytokeratin antibody. Controls were run in the absence of specific antibody. Counter-staining was performed with fluorescent-conjugated goat antimouse IgG, F(ab), fragments.

For histology, distal colon segments from newborn, weanling, and adult rabbits and the tissue remaining after blunt dissection of mucosa was fixed in Cornoy's fluid, embedded in agar, and then dehydrated and embedded in Paraplast (Sherwood Medical Industries, St. Louis, MO) wax. Sections were cut and stained with hematoxylin and eosin.

All photomicroscopy was performed on a Zeiss IM35 microscope using an Olympus OM-2 35-mm camera fitted with a direct mount using Kodak-200 and Kodak Tri-X (Eastman Kodak, Rochester, NY) film. Kodachrome slides of the hematoxylin and eosin sections were prepared, and black and white photographs of these slides were obtained. Therefore photographs shown inFigure 2 represent a negative field image and the hematoxylin staining nuclei appear as bright spots.

Figure 2
figure 2

Histology of newborn, weanling, and adult rabbit distal colon. Hematoxylin and eosin-stained sections of newborn (A), weanling (B), and adult (C) rabbit colon and of newborn rabbit distal colon after dissecting the mucosal layer (D). Kodachrome slides of hematoxylin and eosin-stained sections were photographed, and these black and white illustrations represent negative field images. Bright spots represent nuclei. Original magnification, ×40.

Full size image

Materials. The 24-well plates were from Costar (Cambridge, MA); serum and matrices were obtained from Collaborative Research (Bedford, MA). All other chemicals were obtained from Sigma Chemical Co. (St. Louis, MO) with the following exceptions. The medium was obtained from Hazelton (Lenexa, KS), and the cytokeratin antibody was a generous gift of Dr. Gail A. Hecht(Department of Medicine, University of Illinois at Chicago).

Data analysis (statistics). All data are given as mean ± SEM where n represents the number of animals (pooled groups in the case of newborn animals and weanlings). Comparisons between different groups were performed by calculating the one-tailed ANOVA, whereas comparisons between two groups were performed using paired or unpaired t tests when appropriate. Significance was defined as p < 0.05.

RESULTS

Cell isolation. We had previously demonstrated(17) that a solution comprised of 0.1% Pronase, 0.03% collagenase IV, and 0.07% DTT was optimal for isolation of adult colonocytes. To optimize this enzyme mixture for the isolation of colonocytes from young animals, we compared the efficacy of different concentrations of Pronase, collagenase IV, and DTT in isolating cells from newborn, weanling, and adult rabbits. The cell yield and viability for each solution listed inTable 1 was determined. In all these solutions we obtained a mixture of single cells and clusters of cells as reported earlier for adult colonocytes(17). Aliquots of cells were triturated to break up the clusters before assessing viability. As shown inTable 1, the maximum number of cells and the highest percentage of viability were obtained with solution III (0.05% Pronase, 0.015% collagenase IV, and 0.023% DTT) (p < 0.01) for both newborn and weanlings. In contrast, solution I was best for isolating adult colonocytes as shown previously (Table 1)(17). To compare the cell yield among the newborn, weanlings, and adults, we calculated the number of cells obtained per g of mucosa. As shown inFigure 1, the cell yield is the highest in adult (32± 1.55 × 106) and the lowest in the newborn (9.5 ± 0.5 × 106).

Figure 1
figure 1

Colonocyte yield from mucosa of different age groups. Cell yield per g mucosa of newborn, weanling, and adult rabbit colon isolated under optimal isolation conditions. Values represent mean ± SEM fromn = 4 experiments.

Full size image

Histology and determination of epithelial origin. To determine whether the age-related differences in cell yield can be related to any differences in the colonic mucosa, hematoxylin and eosin-stained sections of the distal colon of the three different groups were compared. Shown inFigure 2 are negative images of hematoxylin and eosin sections (see “Methods”), all at the same magnification. All three tissues are morphologically similar with the expected difference in size. However, although the newborn (Fig. 2A) and weanling(Fig. 2B) colonic mucosa have clearly defined crypts, they are smaller in depth compared with that of the adults(Fig. 2C). Thus the crypt:surface ratio appears to increase with age. Figure 2D also shows that the dissection process removes the mucosa while leaving the muscle layers intact in the newborn. Similar results were observed with dissections in the other two groups (data not shown). To determine whether all of the isolated cells were epithelial in origin, colonocytes obtained from newborn(Fig. 3A), and weanling (Fig. 3C) rabbits were subjected to immunofluorescent staining with a polycolonal antibody against cytokeratins. All of the cells showed a positive staining. In contrast, in the absence of the cytokeratin antibody, no specific labeling of the colonocytes was seen (Fig. 3,B andD).

Figure 3
figure 3

Cytokeratin staining of newborn and weanling rabbit colonocytes after 24 h in culture. Colonocytes from newborn (A andB), and weanling (C and D) rabbits were incubated with a specific mouse primary antibody (A and C) or a nonspecific mouse primary antibody (B and D). All cells were then incubated with a secondary, fluorescein-labeled goat anti-mouse IgG polyclonal antibody. The positive reaction of all of the cells, only in the presence of the anti-cytokeratin antibody (A andC), confirms their epithelial nature. Original magnification,×1200). Note: Photographs A and C were taken with a 1-s time exposure, whereas B and D were exposed for 2 s.

Full size image

Primary culture and viability. There is no significant difference in viability among newborn, weanling, and adult cells for up to 24 h in primary culture when grown on plastic or Matrigel (Fig. 4A). However, there is a gradual, significant (p < 0.01) decline in viability in adult cells when compared with that of newborn and weanling rabbits from 24 to 72 h (Fig. 4A). Growth and proliferation were assessed by performing cell count, DNA and protein assay. The results shown in Figure 4 represent both attached and nonattached cells in culture (also see below). As shown inFigure 4,B-D, cell count, protein and DNA content of newborn (square symbols) and weanling (triangular symbols) colonocytes increased significantly over a 72-h period (p < 0.005 compared with 0-h values) suggesting growth and proliferation in culture. In comparison, adult colonocytes (circular symbols) showed a significantly lower rate of increase (p < 0.01, ANOVA, adultversus newborn or weanlings) in these three parameters over the same time period (Fig. 4,B-D). It must be noted that, although small, adult colonocytes show a significant, ≈10%, increase in cell number and protein and DNA at 48 h compared with that at 0 h (p< 0.02).

Figure 4
figure 4

Growth characteristics of colonocytes in primary culture on plastic (- - - - -) and Matrigel (-). Changes in percent viability(A), cell count (B), protein (mg) (C), and DNA(μg) (D) over 72 h per 5 × 105 plated colonocytes obtained from newborn (▪), weanling (▴), and adult (•) rabbits. Each data point represents mean ± SEM of n = 4 experiments, with each value being run in triplicate. The changes in cell count, protein, and DNA are significantly different at 48 and 72 hvs 0 h at p < 0.005 in newborn and weanling and atp < 0.01 in adult colonocytes. In B-D, newborn and weanling colonocytes plated on Matrigel are significantly (p < 0.01) different from those on plastic at 48 and 72 h. Adult colonocytes grown on Matrigel vs plastic are not different in any parameter except protein at 72 h (p < 0.05).

Full size image

Attachment of colonocytes to different matrices and time course of attachment. One factor that may contribute to the maintenance of colonocytes in culture is the matrix on which they are grown. We systematically evaluated colonocyte attachment to a variety of commercially available matrices and examined the time course of attachment. Attachment was assessed as the percent of total cells adhering to the matrix after 24 h in culture. Attached cells were defined as those cells that required trypsin treatment to be dislodged from the matrix.

As shown in Figure 5, in all age groups the best attachment was observed to Matrigel compared with collagen I, collagen IV, laminin, fibronectin, or plastic. To determine whether the rates of attachment differed during development, the time course of attachment of colonocytes to Matrigel was examined. As shown in Figure 6, although the percent of attachment varied, the rates of attachment in newborn, weanling, and adult colonocytes appeared to plateau by 24 h and were unchanged at 48 and 72 h.

Figure 5
figure 5

Colonocyte attachment to different matrices. Newborn(stippled bars), weanling (hatched bars), and adult(open bars) colonocyte attachment to plastic (P), Matrigel(M), collagen type I (CI), collagen type IV(CIV), laminin (L), and fibronectin (F). Attachment was significantly greater to Matrigel and collagen IV than to the other matrices. Attachment of newborn and weanling colonocytes to Matrigel was significantly greater than that of adult (p < 0.01, ANOVA). Each condition represents the mean ± SEM of four experiments with duplicate samples in each experiment.

Full size image
Figure 6
figure 6

Time course of colonocyte attachment on Matrigel of different age groups. Colonocytes isolated as detailed in“Methods” were plated in quadruplicate and harvested using 0.1% trypsin at 2, 4, 16, 24, 48, and 72 h. Each data point represents mean± SEM of n = 4 experiments. At all time points, values are significantly different in newborn vs weanling vs adult cells, p < 0.01, ANOVA.

Full size image

Efficiency of attachment to various matrices as a function of developmental age. Although colonocytes from all age groups showed the best attachment to Matrigel and the poorest attachment to plastic, there were dramatic age-related differences in the efficiency of attachment. For example, although newborn colonocytes showed an ≈88% attachment, and weanling colonocytes a ≈78% attachment, adult colonocytes showed only a 45% attachment to Matrigel. Similarly, although newborn and weanling colonocytes showed ≈88 and ≈78% attachment to collagen IV, respectively, adult colonocytes showed only a 15% attachment to this matrix. When colonocyte attachment to Matrigel was compared with plastic, the ratio of attachment was 4.1:1 in the newborn (p < 0.01), 5.9:1 in the weanling(p < 0.01), and 9.5:1 in the adult (p < 0.001) animals. The progressive increase in the relative attachment to Matrigel from newborn animals to weanlings to adults is statistically significant (4.1 < 5.9 < 9.5; p < 0.001).

Effect of matrices on growth and proliferation. To determine whether the matrices influence growth and proliferation of colonocytes, cells from the different age groups were plated on plastic, Matrigel, or collagen IV. At 24, 48, and 72 h postplating, attached and nonattached cells were harvested as described above and assessed for viability, cell number, and DNA and protein content. As shown in Figure 4,B-D(solid lines), Matrigel increased (p < 0.01, ANOVA) the cell number, DNA, and protein of newborn and weanling colonocytes by 15% at 48 and 72 h. In contrast it neither affected the proliferation(Fig. 4,B andD) of adult colonocytes nor the viability of any age group (Fig. 4A). The only significant effect of Matrigel on adult colonocytes was a 20% increase in protein content at 72 h(Fig. 4C). Collagen IV had similar effects to Matrigel, there being no statistical difference between the effects of these two matrices (Table 2).

Table 2 Growth characteristics of crypt colonocytes in primary culture plated on different matrices plastic (P), Matrigel (M) and collagen IV (CIV)
Full size table

DISCUSSION

This study is one of the first to define a model for studying colonic epithelial cells, obtained from different stages of development, in culture. An enzymatic digestion procedure, previously shown to be optimal for the isolation of adult colonocytes(17), needed to be modified with respect to composition and time of exposure. This mixture, solution I (0.1% Pronase, 0.03% collagenase, and 0.07% DTT) with an exposure time of 90 min, was optimal for adult colonocytes (Table 1), but proved to be too vigorous for younger animals, yielding a cell population of <70% viability. The yield in adults using solution I compares favorably with the numbers of colonocytes reported by Benya et al.(17) and Kaunitz(19). When the concentrations of the mixture components were reduced to 0.05% Pronase, 0.015% collagenase, and 0.023% DTT (solution III), with a 60-min exposure, >94% viable cells were obtained from the colons of weanlings as well as suckling rabbits. In contrast, the solution III composition and time of exposure was suboptimal for the isolation of adult colonocytes. As shown inFigure 4A, the cells obtained by solution III digestion continued to show ≳94% viability up to 72 h in culture and in preliminary experiments were found to be growing up to 96 h in culture. However, adult colonocytes isolated with solution I showed a decline in viability over 72 h(Fig. 4A). The relative amounts of DTT and enzyme were critical because very low concentrations resulted in low yield and poor viability (Table 1, solution IV). The low yield is probably due to poor dissociation and the low viability due to ineffective breakdown of mucus by DTT.

The isolated cells appear to be epithelial in origin as indicated by their positive immunostaining with anti-cytokeratin antibodies(Fig. 3,A andC). This is similar to our earlier findings with adult rabbit(17) and human(29) colonocytes. As reported for other systems, in the younger animals, the epithelium of the colon appears to have a greater surface cell:crypt cell ratio than does the adult. This is exemplified both in the histologic data presented in Figure 2 and in the cell yield data presented in Figure 1. These findings support our earlier observation that the isolation procedure, employing low speed centrifugation, enriches for crypts(17). The mucosal scrapings used for isolation are comprised of epithelial sheets (data not shown) with no underlying muscle (Fig. 2D). Therefore the differences in cell yield at various developmental stages are due to the relative abundance of crypt cells rather than to differences in the amount of noncellular (e.g. basement membrane) or nonepithelial material in the mucosal sheets.

As shown in Figures 4,B-D, there is also a difference in the rate of growth of the colonocytes in culture. Both the suckling and weanling colonocytes show a near doubling (≈80% increase) of cell number over a 48-h time period, between 24 and 72 h postplating(Fig. 4B). In contrast, the adult colonocytes show a very poor rate, ≈10%, of increase in cell number over the same period of time. Parallel with the increase in cell number there is a ≈2-fold increase in total cell protein content and a >2-fold increase in the DNA content in the neonatal and weanling colonocytes. In marked contrast, there is relatively little increase (≈10%) in both the protein and DNA contents of the adult colonocytes. The data on cell number and DNA content strongly suggest that, whereas the suckling and weanling colonocytes are undergoing proliferation with a doubling time of ≈48 h, the adult colonocytes are undergoing slow, if any, proliferation. A number of factors, singly or in combination, could contribute to these observations. Although the number of stem cells/crypt may not vary with age, it is conceivable that their rates of proliferation may differ with age. Alternatively, because the number of cells/crypt increases with age (Fig. 2), in the face of a constant number of stem cells/crypt, the proportion of stem cells to total cells in the crypt-enriched population plated is lower in the adult. This could partly contribute to the observed lower rate of proliferation in the adult. Another explanation is that the culture conditions used are more conducive to growth in colonocytes at early stages of development. In addition, as suggested by Gordon(30), the stem cells of young animals could give rise to multipotent cells which are more conducive to proliferation underin vitro conditions than the unipotent cells that arise from adult isolates. Whatever the cause(s), the results highlight that there are age-related differences in the proliferative capacity of colonocytes.

The developmental differences in colonocyte function are also highlighted in the ability of these cells to attach to various matrices. As demonstrated for other species, including normal and colon carcinoma-derived human colorectal samples(31), rabbit colonocytes attached preferentially to collagen IV and complex matrices like Matrigel(Fig. 5). Matrigel is a basement membrane-derived extracellular matrix, composed primarily of collagen IV, laminin, heparin sulfate proteoglycans, and entactin(32), which may explain the preferential attachment of colonocytes to this matrix. As in the case of proliferation, the attachment capabilities of colonocytes to all matrices declines as the age of the animal advances. With the exception of a plastic substratum, neonatal and weanling colonocytes attached to all biomatrices with an efficiency of >50%. In contrast, the adult colonocytes showed only a ≈45% attachment to Matrigel and <25% attachment to the other matrices. The reason for these differences is not known, but we speculate that differences must be related to the stage of growth and differentiation of these colonocytes. It is conceivable that there may be a greater percentage of undifferentiated cells that are more flexible in their requirements for attachment in the cell population isolated from younger animals. This is supported by our observation that the relative attachment of colonocytes to complex matrices compared with plastic is enhanced at later stages of development. Clearly the different cell types express different surface proteins, including receptors for extracellular matrix proteins. The nature of the proteins and these interactions need to be delineated.

Our results imply that the link between attachment and proliferation is indirect at best. First, at all three stages attachment had plateaued by 24 h postplating (Fig. 6), and the cells continued to proliferate beyond this time. Thus, the 15% increase in proliferation of neonatal and weanling colonocytes grown on Matrigel and collagen IV, over those grown on plastic, cannot account for the >2-fold increase in cell numbers at 48 h. Second, there appears to be no strict correlation between the number of cells and the percent of attachment at 24 h; i.e. both the attached and nonattached cells appear to proliferate. For example, 5 × 105 adult colonocytes plated on plastic or on Matrigel resulted in 5.7× 105 and 5.8 × 105 cells at 24 h, respectively. This was despite the fact that only 4.5% of the cells showed attachment to plastic, whereas 45.3% of the cells attached to Matrigel. It is of interest that, at 24 h, weanling and newborn colonocytes also show similar ≈20% increases in cell number irrespective of whether they are plated on Matrigel(75 and 88% attachment) or on plastic (14 and 22% attachment). Thus, at 24 h, attachment did not have a profound influence on rates of proliferation.

In summary, we have developed an in vitro model for studying development-related changes in the colonic epithelium. Our studies demonstrate that neonatal and weanling colonocytes can attach to extracellular matrices such as collagen IV and Matrigel with a >80% efficiency. The marked differences between the proliferative rates and their ability to attach to biomatrices support studies in intact tissues where developmental differences in colonic epithelial function including ion transport have been reported. Our long-term goal is to identify growth factors and extracellular matrix components that are required to maintain cell proliferation and differentiation of the colonic epithelium in vitro. In addition, this model will be a useful tool to decipher the cellular and molecular changes occurring in colonic epithelia during development.