Caco-2 cells can serve as an in vitro model to study the enterocyte transport of lipids, and lipid-soluble drugs/vitamins. The permeable membrane system separates the apical from the basolateral compartment, while the lentivirus expression system offers an effective gene overexpression method. The isolation of lipoproteins is confirmed by TEM.
Studies of dietary fat absorption are generally conducted by using an animal model equipped with a lymph cannula. Although this animal model is widely accepted as the in vivo model of dietary fat absorption, the surgical techniques involved are challenging and expensive. Genetic manipulation of the animal model is also costly and time consuming. The alternative in vitro model is arguably more affordable, timesaving, and less challenging. Importantly, the in vitro model allows investigators to examine the enterocytes as an isolated system, reducing the complexity inherent in the whole organism model. This paper describes how human colon carcinoma cells (Caco-2) can serve as an in vitro model to study the enterocyte transport of lipids, and lipid-soluble drugs and vitamins. It explains the proper maintenance of Caco-2 cells and the preparation of their lipid mixture; and it further discusses the valuable option of using the permeable membrane system. Since differentiated Caco-2 cells are polarized, the main advantage of using the permeable membrane system is that it separates the apical from the basolateral compartment. Consequently, the lipid mixture can be added to the apical compartment while the lipoproteins can be collected from the basolateral compartment. In addition, the effectiveness of the lentivirus expression system in upregulating gene expression in Caco-2 cells is discussed. Lastly, this paper describes how to confirm the successful isolation of intestinal lipoproteins by transmission electron microscopy (TEM).
Studies of intestinal absorption of dietary fat, and lipid-soluble drugs and vitamins can be conducted in vivo by using a lymph fistula model 1–4. However, the surgical techniques involved are not only challenging, but also costly. Although in vivo approaches based on fecal analysis may be utilized, they are used mainly to determine the percent uptake by the gastrointestinal tract 2,5. The in vitro model described in this paper is more cost effective, and the techniques involved are arguably less challenging. Genetic modification studies are also more economical and less time-consuming when they are conducted using this in vitro model.
Since lipid-soluble materials that are taken up by enterocytes are packaged into lipoproteins 6,7, the effectiveness of this in vitro model to produce lipoproteins is crucial. The two main intestinal lipoproteins are chylomicrons and very low-density lipoproteins (VLDL). Chylomicrons, defined as lipoproteins with 80 nm or more in diameter, are produced strictly by the small intestine when lipids are abundantly present in the gastrointestinal lumen. Since they are the largest lipoproteins, chylomicrons are conceivably the most efficient lipid transporters. This in vitro model, which is capable of producing chylomicrons 8, can be used to study dietary fat absorption, lipid-soluble vitamin absorption by the gut, and oral lipophilic drug bioavailability. The presence of lipid-soluble molecules, vitamins, or drugs in the lipoprotein fraction is an indicator of their absorption by the small intestine. As previously discussed, this model can be used to improve oral lipophilic drug bioavailability 6.
This paper describes how Caco-2 cells should be maintained in permeable membrane or regular tissue culture dishes, how the lipid mixture for stimulating lipoprotein production should be prepared, how the lentivirus expression system can be employed to achieve effective overexpression, and how the isolated lipoproteins should be analyzed.
1. Maintenance of the Caco-2 Cells
2. Gene Overexpression
3. Stimulating the Lipoprotein Secretion
4. Lipoprotein Isolation (Figure 3)
5. TEM Analysis
Figure 1 displays normal 13-days post-confluent Caco-2 cells. The appearance of dome-shaped structures and intracellular lipid droplets are characteristic of differentiated Caco-2 cells. When the Caco-2 cells are not dispersed equally during seeding, they will clump and overgrow in certain areas of the dish; and there will be a few areas in the dish without any cells. Swirling and placing the dish on a slanted surface should be avoided. It is also important to note that post-confluent Caco-2 cells are more susceptible to detachment when new growth media is added to the cells too roughly. Therefore, the new media should be added gently to prevent cell detachment.
Based on these studies, the transfection efficiency of Caco-2 cells was between 30-60% (Figure 2 top panels). In contrast, the transduction efficiency of Caco-2 using the lentivirus expression system was approximately 100% (Figure 2 bottom panels). Figure 2 also showed that the optimal amount of concentrated lentivirus was 10 μl. The transduced Caco-2 cells were maintained up to 12 passages. As shown, even after 12 passages the transduced cells still expressed eGFP. The transduction efficiency clearly depends on the lentivirus concentration. The confirmation of the transfection/transduction efficiency is critical, and should be performed as a routine preliminary analysis prior to the actual experiments. Although Western blot analysis can be used to estimate the percent increase of the gene expression, it should not be the primary method to determine the transfection/transduction efficiency.
Using the NaCl density gradient ultracentrifugation method (Figure 3), the lipoproteins secreted by the Caco-2 cells were isolated and analyzed on a TEM. Some of the chylomicrons, lipoproteins larger than 80 nm in diameter, were depicted in Figure 4. The smaller lipoproteins, VLDLs, were also present. It is essential to confirm the successful isolation of both chylomicrons and VLDLs on a TEM. As previously discussed 8, biochemical analysis should not be the primary method of confirmation. The absence of lipoproteins, specifically chylomicrons, indicates that the lipid transport by the Caco-2 cells is not efficient. Consequently, they will not serve as a good model to study lipid transport. The number of chylomicron particles relative to the total number of lipoprotein particles can be counted 8. A high percentage indicates efficient lipid transport.
Figure 1. Representative micrograph of 13-day post-confluent Caco-2 cells. Intracellular lipid droplets are visible in some cells (red arrow). The unique dome-shaped structures (black arrow) formed by some Caco-2 cells are generally present only after the cells have reached confluence. Magnification = 100X. Please click here to view a larger version of this figure.
Figure 2. Comparison of the effectiveness of the lipid-based transfection and the lentivirus expression system. Top panel: Caco-2 cells were transfected with pLL3.7 eGFP using the non-liposomal transfection reagent (Scale bar = 20 μm). Bottom 4 panels: Caco-2 cells were transduced with pLL3.7 eGFP using the lentivirus expression system with varying amounts (1, 2, 5, or 10 μl) of the concentrated lentivirus (LV). The displayed cells were from the 12th passage of the original transduced cells. The left panels show the DAPI staining, the middle panels show the GFP fluorescence, and the right panels show the merged images. The lentivirus expression system was evidently more effective than the lipid-based transfection. Please click here to view a larger version of this figure.
Figure 3. Isolation of intestinal lipoproteins using NaCl density gradient ultracentrifugation. The density of lipoprotein-containing media is adjusted to 1.2 g/ml by adding the appropriate amount of NaCl. The volume of the sample also needs to be adjusted so that it will fill the entire ultracentrifuge tube to prevent collapse. To achieve a density gradient, 0.5 ml of water is overlaid gently on the 1.2 g/ml density solution. The sample is then spun at 429,460 x g for 24 hr using a T865 rotor (equivalent to 2.15 Svedberg units). The intestinal lipoproteins should be immediately recovered by gentle pipetting. Please click here to view a larger version of this figure.
Figure 4. Representative electron micrograph of the lipoproteins produced by the Caco-2 cells. The lipoproteins isolated by NaCl density gradient ultracentrifugation were negatively stained with 2% phosphotungstic acid (pH 6.0). Chylomicrons, lipid particles larger than 80 nm in diameter, and VLDLs, those smaller than 80 nm, are both present. Scale bar = 100 nm. Please click here to view a larger version of this figure.
In this paper, the two systems that can be used to maintain Caco-2 cells are described, namely, the regular tissue culture dish and the permeable membrane. The benefits of using the permeable membrane system include the separation of the apical and the basolateral compartments, and the ability to incubate the lipid mixture and collect the lipoprotein secretion simultaneously. However, the permeable membrane inserts are expensive, and their polycarbonate membrane does not allow for good cell visibility. One of the advantages of this in vitro model is that genetic manipulation studies will be more economical and less time-consuming. For better effectiveness, the lentivirus expression system should be used. The transduced Caco-2 cells generally maintain the expression of their transgene.
The ability of Caco-2 cells to produce chylomicrons is of paramount importance. Without this ability, Caco-2 cells would not be able to efficiently transport lipophilic materials, including but not limited to liphophilic drugs, vitamin A, D, E, and K, and any lipid-soluble nutrients. The proper methods to challenge Caco-2 cells to produce both VLDL particles and chylomicrons are described. The successful isolation of these lipoprotein particles should be confirmed on a TEM. Based on the current literature, this Caco-2 model offers the most efficient lipid transport among other Caco-2 models 12–14. However, the in vivo lymph cannulation model still transports lipids more efficiently than any in vitro model. The underlying reasons have been recently discussed 8, namely because Caco-2 cells produce 2 different isoforms of apolipoprotein B, Caco-2 cells can’t synthesize triglycerides from monoglycerides, and the serum component critical for chylomicron biogenesis may be lower in the growth media. It is also important to realize the limitation of this in vitro model; this in vitro model excludes some potential important factors, such as gut motility, anatomy of the gut, and the interaction with other organ systems.
The main factors that allow Caco-2 cells to produce chylomicrons efficiently are the type/amount of lipids used and cellular differentiation 8. Without the proper combination of these factors, Caco-2 cells will not produce a significant number of chylomicrons 8. Of note, NaCl density gradient ultracentrifugation should be performed properly. The successful isolation of lipoproteins depends on timing (immediate without significant delay), good sample handling (sturdy without much agitation), and careful pipetting (getting only the top layer). Proper technique can be practiced by using pre-stained lipids to help visualize the lipoprotein layer 8. Besides TEM, biochemical analyses, i.e., apolipoprotein B and triglyceride analyses, can also be used to confirm the successful isolation of lipoproteins. These biochemical analyses, which we have previously reported 8, can also serve as methods in quantifying absorption. However, TEM should still be performed due to the tendency of the lipoproteins to aggregate2, causing a potential overestimation of chylomicron production.
This in vitro model is particularly useful for studying dietary fat absorption and intestinal absorption of lipophilic drugs, vitamins, and other lipid-soluble nutrients. It can also serve as a model to improve poor bioavailability of oral lipophilic drugs. Since lipid-soluble materials are packaged into chylomicrons by enterocytes for transport to the circulation, chylomicron-producing Caco-2 cells will be more efficient in absorbing lipophilic drugs. In addition, this model allows investigators to determine the role of a specific gene in drug absorption by the gut (pharmacogenetics). It also allows investigators to compare the effect of various dietary fats on oral lipophilic drug bioavailability. All of these applications have been discussed previously 6.
The authors have nothing to disclose.
This work was supported by the Seed Grant Award from California Northstate University College of Pharmacy (to AMN). The authors would like to thank California Northstate University College of Pharmacy for covering the publication cost of this article, and George Talbott for his help in editing this manuscript.
DMEM | VWR | 16750-112 | Pre-warm the growth media in individual tissue culture dish before adding cells |
FBS | Fisher | 3600511 | We do not heat inactivate our serum |
Trypsin | VWR | 45000-660 | Cells can be washed with PBS prior to trypsin treatment |
Permeable membrane system | Fisher | 7200173 | 10-cm dish, 3-mm pore size, polycarbonate membrane |
10 cm dish | Fisher | 08-772-E | Tissue culture dish |
15 cm dish | Fisher | 08-772-24 | Tissue culture dish |
24 well plate | Fisher | 12565163 | Tissue culture plate |
Lipid-based transfection reagent | Fisher | PRE2691 | Can be substituted with other transfection reagent |
Reduced serum media | Invitrogen | 11058021 | For transfection |
pLL3.7 eGFP | Addgene | 11795 | https://www.addgene.org/11795/ |
Bottle-top filter | Fisher | 9761120 | 0.45 mm pore |
Polybrene | Fisher | NC9840454 | 10 mg/mL |
Oleic acid | Sigma | 01383-5G | Prevent freeze-thaw cycle |
Lecithin | Fisher | IC10214625 | Egg lecithin |
Sodium taurocholate | Fisher | NC9620276 | Product discontinued; alternative catalog number: 50-121-7956 |
Protease inhibitor cocktail tablet (EDTA-free) | Fisher | 5892791001 | Used mainly for samples that need TEM analysis |
Polycarbonate ultracentrifuge tube | Fisher | NC9696153 | Reusing it multiple times will collapse the tube |
Lid for ultracentrifuge tube | Fisher | NC9796914 | A tool is required to remove the tube/lid from rotor |
Syringe | Fisher | 50-949-261 | Disposable |
Syringe filter | Fisher | 09-719C | Pore size = 0.2 mm; nylon |
Phosphotungstic acid | Fisher | AC208310250 | For preparing 2% phosphotungstic acid, pH 6.0 |
Tweezer | Fisher | 50-238-62 | Extra fine and strong tips |
Formvar/carbon grid | Fisher | 50-260-34 | Formvar/carbon film square grid 400 Copper |