The Uptake of Soluble and Particulate Antigens by Epithelial Cells in the Mouse Small Intestine

Intestinal epithelial cells (IECs) overlying the villi play a prominent role in absorption of digested nutrients and establish a barrier that separates the internal milieu from potentially harmful microbial antigens. Several mechanisms by which antigens of dietary and microbial origin enter the body have been identified; however whether IECs play a role in antigen uptake is not known. Using in vivo imaging of the mouse small intestine, we investigated whether epithelial cells (enterocytes) play an active role in the uptake (sampling) of lumen antigens. We found that small molecular weight antigens such as chicken ovalbumin, dextran, and bacterial LPS enter the lamina propria, the loose connective tissue which lies beneath the epithelium via goblet cell associated passageways. However, epithelial cells overlying the villi can internalize particulate antigens such as bacterial cell debris and inert nanoparticles (NPs), which are then found co-localizing with the CD11c+ dendritic cells in the lamina propria. The extent of NP uptake by IECs depends on their size: 20–40 nm NPs are taken up readily, while NPs larger than 100 nm are taken up mainly by the epithelial cells overlying Peyer's patches. Blocking NPs with small proteins or conjugating them with ovalbumin does not inhibit their uptake. However, the uptake of 40 nm NPs can be inhibited when they are administered with an endocytosis inhibitor (chlorpromazine). Delineating the mechanisms of antigen uptake in the gut is essential for understanding how tolerance and immunity to lumen antigens are generated, and for the development of mucosal vaccines and therapies.


Introduction
The mucosa of the gastro-intestinal tract is continuously exposed to dietary and microbial antigens. As an interface between the outside environment (lumen) and the inner body, gut-associated lymphoid tissue (GALT) maintains a delicate balance of inducing immunity against pathogens and tolerance to the antigens originating from the diet and intestinal microflora [1,2,3]. Among other factors, the route of antigen uptake and the nature of the antigen dictate the ensuing immune responses in the deeper lymphoid tissues. Lymphoid tissues of the small intestine (SI), such as Peyer's patches, contain M cells that take up large antigens (bacteria, particles, etc.) and deliver them to the underlying immune cells to initiate immune responses [4]. Dendritic cells (DCs) of the SI lamina propria (LP), namely CD11c+ [5], CD103+ [6] and CX3CR1+ [7] DCs extend their dendrites between the IECs and sample lumen antigens. Goblet cell-associated passageways (GAPs) allow the entry of soluble protein antigens, but not inert particles (0.02-2 mm) into the LP [8]. Epithelial cells overlying the GALT and the LP represent a physical barrier that separates the body from the lumen and are also the first point of contact between the host immune system and lumen antigens. Conventional IECs (enterocytes) absorb digested, small molecular weight dietary antigens via the transcellular pathway [9,10,11] and allow small molecules and inert experimental probes (5-10 Å ) to access the LP via the tight junctions (paracellular pathway) [9,12,13]. IECs differ from M cells in that they contain closely packed microvilli [14] and express 400-500 nm-thick mucin-like glycoproteins that form a layer covering the tips of the microvilli; whereas M cells lack microvilli, do not secrete mucus and generally lack the thick glycocalyx layer [15,16,17]. The mucus layer traps microorganisms and other large inert antigens, preventing their direct contact with the IEC membranes [2,18,19] and access to inter-microvillar endocytic domains [15]. However, smaller pathogens such as norovirus (20-30 nm) and human papilloma virus (,55 nm) can readily diffuse through cervical mucus [20] which is similar in physical properties to the mucus covering the IECs [21]. Whether IECs play an active role in the uptake and sampling of small particulate lumen antigens such as viruses and bacterial cell debris in vivo is not known. Mathiowitz and coworkers showed that polymer particles 40-120 nm in size, engineered to display strong adhesive interactions with mucus and cell membranes are taken up by IECs and facilitate the transport of conjugated substances into the LP [22]. In contrast, larger size polystyrene and poly(lactic acid) particles are taken up exclusively by M cells associated with Peyer's patches [23,24,25]. Here we used fluorescently labeled antigens and polystyrene NPs to examine their in vivo uptake by confocal and immunofluorescence microscopy (IFM). We report that routes of antigen uptake depend on the nature of the antigen. Soluble, small molecular weight antigens enter the LP via GAPs, while 20 and 40 nm NPs cross the mucus layer and are internalized by the IECs of the villi. Internalized NPs are then found in the underlying CD11c+ LP DCs, blood, and lymphatic ducts of the villi through which they reach the mesenteric lymph nodes (MLNs).

Ethics statement
All animal experiments were performed in accordance with NIH guidelines, the Animal Welfare Act, and US federal law. Animals were housed in centralized AAALAC-accredited research animal facilities, staffed with trained husbandry, technical, and veterinary personnel. All experiments were conducted in accordance with protocols approved by the Southern Illinois University Institutional Animal Care and Use Committee and Saint Louis University Animal Studies Committee.

Administration of antigens to the mice
For short-term experiments 50-100 ml of PBS containing ether NPs (1-10%), E.coli particles (0.1-1 mg) and/or soluble antigens (dextran-fluorescein (0.5-2 mg), Ova-fluorescein (0.2-2 mg), LPS-Alexa FluorH 488 (15-20 mg) were administered directly into the SI lumen of anesthetized mice, no further than 10 cm upstream of the ileo-cecal junction. For this, mice were fasted for 2 hours then anesthetized with isoflurane delivered in a stream of oxygen. The SI was exposed via a small incision in the abdominal wall and antigens were injected using a 30 g needle. During anesthesia animals were placed on a warming pad and covered with a cloth to maintain a stable body temperature. In some experiments antigens were administered in a 200 ml volume of PBS via a gastric gavage using a round-tip needle. Before gastric gavage, mice were fasted for 2-3 hours.
In vivo imaging of the SI At pre-determined times after antigen administration to the SI (or via a gastric gavage), a small incision was made in the wall of the SI. The lumen side of the SI was secured under a cover glass with a thin film of Vetbond (Ted Pella). To inhibit intestinal peristalsis mice were injected sub-cutaneously with 200 ml of PBS containing 1 mg/ml scopolamine 15-30 minutes before imaging.
The mouse was then placed in a custom-made restraint device on the microscope stage of an Olympus FV1000 MPE scanning confocal microscope and imaged using a 206 0.95 NA water immersion objective (Olympus). Fluorophores were excited using a Chameleon UltraII laser (Coherent) with excitation wavelengths from 640 to 1080 nm. Ten to 50 sequential Z-stacks were acquired at ,1-2 mm spacing, rendered into 3D images using Volocity software (Perkin Elmer) and examined for the presence of fluorescent particles in IECs, LP, and the MLNs. In some experiments dextran-fluorescein was injected i.v. in a tail vein 30 min before in vivo imaging to highlight blood and lymphatic vessels in the SI.

Analysis of particle internalization by immunofluorescent microscopy (IFM)
Parallel experiments were performed for analysis of NP internalization and location within the tissues by IFM. At different times after NP administration (per-orally or in the SI) mice were euthanized and sections of the SI were excised and snap-frozen in Tissue-TekH O.C.T. freezing compound on dry ice. Tissue cryosections (5-7 mm thick) were fixed in 4% paraformaldehyde (PFA), washed with PBS then incubated with blocking buffer (Thermo Scientific) for 10-15 minutes. After blocking, sections were stained with antibodies or phalloidin and imaged with a Leica DM4000B fluorescent microscope. Images were analyzed with Volocity software. For TEM imaging sections of the SI were prepared as described previously [26].

Quantification of NPs accumulated in the MLNs
In these experiments 40 nm NPs were administered to the lumen of the SI with or without CPZ. Forty minutes after NP administration mice were euthanized, MLNs were excised, and draining lymphatic ducts and the adipose tissue surrounding the MLNs were removed under a dissecting microscope. Fluorescence intensity of the isolated MLNs was measured with a Shimadzu RF-5301 PC spectrofluorophotometer. Diluted 40 nm NPs in PBS were used to set a calibration curve for the instrument.

Quantification of NPs accumulated in the MLNs using Volocity
Experiments were conducted as described in the above paragraph. Excised MLNs were snap-frozen in OCT on dry ice then 7 mm sections of the MLNs were imaged with a Leica DM4000B fluorescent microscope at 6306 magnification. Acquired images from all MLN compartments were analyzed with Volocity software. Care was taken to quantify pixels specific for fluorescent NPs only within the MLN tissue. The number of pixels specific for fluorescent NPs was expressed as the percentage of pixels per image (MLN tissue surface area). Data acquired from 10 representative images per mouse taken at 6306 magnification from MLN compartments (high or low NP concentration) were used for statistical analysis.
Conjugation of particles with ovalbumin n a 2 mL eppendorf tube 200 ml of NPs (2 wt % solid solution) were mixed with 800 mL of 100 mM PBS solution and 8.0 mg of Ova (Sigma). The mixture was incubated at room temperature for 15 minutes and then 8.0 mg of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide-hydrochloride (EDAC-HCl) was added. The reaction mixture was incubated for 2 hours at room temperature with regular agitation. The resulting particle dispersion was then dialyzed using a Float-A-LyzerH membrane (100 kDa cutoff) for 3 days against 100 mM PBS (pH 7.4) that was changed daily. The resulting protein-conjugated NPs were diluted in PBS and stored at 4uC until used.

Analysis of NPs conjugated to OVA immobilized on a nylon membrane
To confirm that the NPs were conjugated to Ova, NPs, Ova, or Ova-conjugated NPs (NP-Ova) were spotted on a 0.2 mm nylon membrane. The membrane was washed with Tris-buffered solution containing 0.1% Tween 20 (Sigma) (TTBS), blocked with TTBS containing 5% skim milk (Difco) for 1 hour then incubated with biotinylated rabbit anti-Ova antibodies (Thermo Scientific) for 1 hour. After 3 washes the membrane was incubated with streptavidin-FITC (Biolegend) for 1 hour, washed, dried then imaged with a Leica DM4000B fluorescent microscope using a 2.56 objective.

Isolation of epithelial cells from the SI
Sections of the SI were flushed with cold PBS and the Peyer's patches were excised and discarded. The SI was cut longitudinally, then into 2 cm sections and placed in 250 ml Erlenmeyer flasks containing 40 ml of pre-warmed (37uC) HBSS (HyClone) supplemented with 5 mM EDTA. Flasks were incubated for 15 minutes on a shaking rotator at 37uC. Supernatant containing IECs was passed through a 70 mm cell strainer (BD Falcon) and IECs were pelleted by centrifugation. Pelleted IECs were washed 3 times with RPMI-10 (HyClone) supplemented with penicillin (100 u/ml) and streptomycin (100 mg/ml). To confirm that isolated cells were IECs they were fixed in 4% PFA, washed with PBS then incubated with mouse anti-E-cadherin antibodies for 1 hour. IECs were washed in PBS then incubated with FITCconjugated goat anti-mouse antibodies for 1 hour. Isolated IECs were placed on a glass slide and imaged with a fluorescent microscope. A combination of the same antibodies was used to confirm the expression of E-cadherin in isolated IECs by western blot analysis. As a negative control for E-cadherin expression, isolated spleen lymphocytes were used.

Statistical analysis
Each experiment was repeated at least 3 times. Data were analyzed using procedures of SAS software. Group means were separated by using Student's t-test and were considered significantly different at P,0.05. Data are expressed as mean 6 standard deviation (SD) of the mean.

Entry of small molecular weight antigens into the LP via GAPs in vivo
In vivo imaging allowed us to examine the uptake and distribution of antigens in the SI in a living mouse with an intact epithelial barrier and uninterrupted blood and lymphatic flow. The routes of antigen administration and antigen uptake are summarized in Table 1. As expected, the small molecular weight dextran (40 kDa) and Ova (45 kDa) entered the GAPs and the LP when injected into the intestinal lumen ( Figure 1A (yellow arrow)) [8]. Salmonella LPS (3 kDa) also entered via GAPs ( Figure 1B (pink arrows, inset)) and accumulated in the LP of the villi ( Figure 1B (circled)). We then sought to examine whether larger particulate bacterial antigens enter the LP via GAPs by using fixed, fluorescent E.coli as a model antigen. Cell debris, generated by sonicating E. coli (but not intact bacteria) entered the LP of the villi ( Figure 1C (blue arrows)). We did not observe any significant amount of E.coli cell debris entering the LP via GAPs, indicating that the antigen size was important for their entry into the LP via GAPs.  Table 1. Routes of uptake (entry) of soluble and particulate antigens in the small intestine (SI) of mice.

Mode of administration GAPs IECs
Peyer's patches To examine how the particle size might affect their uptake, in further studies we used negatively charged fluorescent polystyrene particles ranging in size from 20 nm to 2 mm. These particles are available in a variety of sizes, do not bind to cells, and are impregnated with fluorophores, which makes them suitable for in vivo imaging and IFM analysis. To examine whether NPs enter the LP via GAPs we co-administered NPs with either dextran or Ova per-orally. This approach enabled us to simultaneously examine the entry of NPs and fluorescent soluble antigens into the LP. Within the time examined (30 minutes to 1 hour after   Figure S1 A, B). These structures become apparent when highlighted by fluorescent antigens. Using TEM we observed that NPs were lodged between the microvilli (Figure S2 A, B) of the IECs, and did not appear not to be uniform in size (Figure S2 B). This could be because the specified sizes of NPs available commercially represent their nominal bead diameters. However, due to their very small size, 20 and 40 nm NPs are less uniform in size when manufactured, with a higher coefficient of variation (up to 20% (6 nominal diameter)), Invitrogen). Because of batch variation during manufacturing it is possible that occasional larger particle is found in the mix of NPs. Within the epithelial cells NPs were found in close proximity to the DAPI-stained nuclei of IECs  Internalization of NPs can be inhibited by an endocytosis inhibitor Using TEM we did observe 40 nm NPs lodged between the microvilli of IECs ( Figure S2), however due to a lack of contrast within the tissues we had difficulties visualizing NPs when tissues were prepared using standard fixation techniques. This made it difficult to unequivocally identify the endocytic compartments harboring NPs within the IEC cytoplasm. We then administered NPs with either genistein (an inhibitor of caveolae-mediated endocytosis) or CPZ (an inhibitor of clathrin-mediated endocytosis) into the SI and visualized the NP uptake in vivo. In our hands, co-administration of genistein with 20 or 40 nm NPs did not inhibit their uptake by IECs, even when genistein was used at 1 mM, a 5-fold higher concentration than is commonly used for cultured cells (not shown). In contrast to this, co-administration with CPZ did greatly inhibit the uptake of 40 nm NPs by IECs in vivo, even though the amount of CPZ used was 5-fold higher (50 mg/ml) than what is typically used for in vitro studies (Figure 6  A, B). CPZ did not affect the average numbers of GAPs per villi ( Figure 6 F) or the dextran entry into the LP via GAPs (Figure 6 C-E, G). We then used 10-fold higher concentration of CPZ, however similarly to previous studies in which lower CPZ concentration was used (Figure 6 A, B), the NP uptake was not inhibited in all areas of the SI examined in vivo. This finding could be due to the fact that NPs were administered in a small volume of PBS (50-100 ml) and that the in vivo absorption of CPZ in some areas may be more extensive than in others.

Quantification of in vivo NP uptake with and without CPZ inhibition
Parallel experiments were conducted in which 40 nm NPs with or without CPZ were administered to the SI of the anesthetized mice. Forty minutes after NP administration MLNs were either frozen for IFM analysis or were analyzed by spectrofluorophotometry. For IFM analysis, cryosections of entire MLNs were imaged at 6306 and the amount of NPs (red pixels) was quanified within the MLN tissue. MLNs of mice that received CPZ had significantly less NPs in regions of high NP concentration (Figure 7 A-C), as well as in MLN capsules (Figure 7 D-F). This finding was confirmed with spectrofluorophotometry ( Table 2). For measuring fluorescence intensity in MLNs, we generated a calibration curve using known concentrations of NPs. In all three experiments we observed a lower intensity of fluorescence (U) in MLNs isolated from mice which were administered NPs with CPZ compared to the intensity of fluorescence in MLNs of mice to which only NPs were administered (Table 2).

NP size affects their internalization
Regardless of the route of administration (per-oral or injected in the SI), the uptake of smaller particles was more efficient. Peyer's patches internalized a significant amount of NPs (20 nm-100 nm) ( Figure S3 A, B), but very few large particles (0.5-2 mm) within the same time frame (not shown). When 20 nm and 40 nm NPs were given per-orally they reached the serosa of the SI in large amounts ( Figure S3 C). In contrast, when 40 and 1000 nm NPs were administered, large amount of 40 nm NPs (green), but very few 1000 nm NPs (red) reached the serosa of the SI (Figure S3 D,  (arrows)). For the most part, the larger NPs adhered to the mucus and made less direct contact with the epithelial cells overlying the Peyer's patches (not shown) and the villi (Figure S3 E). When 100 nm or 500 nm NPs were administered into the SI they were predominantly found in the lumen and were not internalized by IECs (Figure S4 B-E).

NPs conjugated to Ova are internalized efficiently and transport Ova to the deeper lymphoid tissues
We then examined whether coating NPs with a protein antigen altered their uptake. For this, we conjugated 20 nm NPs with Ova and immobilized conjugated NPs, NPs alone, or Ova onto a nylon membrane. We probed the membrane with anti-Ova antibodies to confirm co-localization of NPs with Ova. Blot dots in which Ova or NPs alone were spotted produced a signal only in the green or the red channels respectively (Figure S5 A, top and middle rows). Dots in which NP-Ova were spotted produced signal in both the green and the red channels and the two colors co-localized perfectly (Figure S5 A, bottom row). Conjugating 20 nm NPs to Ova did not inhibit their uptake (Figure S5 B, C), moreover Ova and NPs were detected co-localizing 30 minutes after administration in the lumen of the SI and in circulation of the SI on the serosal side ( Figure S5 D-F, (white arrows)).

Discussion
Several routes of antigen uptake from the intestinal lumen have been reported and it is becoming apparent that the mode of antigen uptake plays a role in ensuing immune responses. Whether IECs (enterocytes) play a role in sampling lumen antigens in vivo is not known, although cultured IECs can internalize bacteria, flagellin, peptidoglycan, LPS, and particles of various sizes [27,28]. Here we have used in vivo imaging of the SI by two-photon/ confocal microscopy and IFM of cryosections to examine whether IECs internalize particulate lumen antigens such as viruses, bacterial cell debris, and dietary particles. Initially we used fluorescent E. coli as a model particulate antigen and did not observe any appreciable uptake by cells of the epithelium or a significant accumulation of E.coli in the LP of the villi. However, E.coli cell debris produced by sonication did accumulate in the LP of the villi, but it did not appear to enter via GAPs, indicating that the size of particulate antigen was important for its internalization via GAPs. Sonication of E. coli produces cell debris consisting of particulate antigens of varying sizes, making it impossible to examine whether and how the particle size affects their internalization in vivo. Thus, for further studies we used fluorescent particles of known sizes which do not exhibit any endogenous binding affinity to IECs. We found that the uptake of NPs in the SI and their transport was rapid: within 30 minutes of administration to the SI, NPs could be observed in the LP, within the serosal circulation, and in the lymphatics of the SI. In the LP of the villi, NPs co-localized with CD11c+ DCs, although in cryosections of the SI we did not observe any instance of a direct uptake of lumen NPs by CD11c+ DCs of the villi. More recently it was reported that LP CD103+ CD11c+ DCs take up bacterial antigens from the lumen of the SI. We consistently observed NPs located on the basolateral side of the IECs in vivo and using IFM found that a very small number of CD103+ DCs populate the LP of the villi in a steady-state, leading us to consider that IECs could play a role in NP internalization. Although we do find NPs colocalizing with CD11c+ DCs in the LP, we cannot exclude the possibility that some NPs may enter the interstitial spaces, then the blood or lymphatic ducts of the villi, enabling their rapid transport to the MLNs. We found that in addition to Ova and dextran, Salmonella LPS enters the LP mainly via GAPs, indicating that in steady-state GAPs may play a role in internalization of small microbial antigens originating from microflora. In vivo imaging of IECs highlighted by NP fluorescence revealed a network of conduit-like structures on the surface of the villi and Peyer's patches that appear to guide soluble antigens into the GAPs. These structures become apparent when highlighted by fluorescent soluble antigens for in vivo imaging, but are not readily discernible in cryosections of the SI. Conduit-like structures do not represent paracellular leakage of the antigen, since they are on the surface of the villi and almost always reach the LP via GAPs. Paracellular  leakage of antigen is observed in deceased animals, when the tight junctions lose integrity and the antigen leaks into the LP. In that scenario leakage of fluorescein-labeled dextran appears as fine green lines of fluorescence between all IECs, which is not observed when the SI is imaged in a living animal. Since conduit-like structures are found almost always connected to GAPs, we speculate that they may play a role in directing small soluble antigen to GAPs or alternatively, they might serve to transport/ distribute mucus on the surface of the epithelium. The internalization of NPs was most pronounced in the upper 3 rd of the villi, especially at the tips of the villi which appear as red patches due to the presence of NPs. The presence of mucus in the intestines makes it difficult for dietary particles, resident microflora, and their products to contact the surface of the IEC. We found that 20 and 40 nm NPs cross the mucus layer and make direct contact with IECs. Our finding is in agreement with reports that capsid virus-like particles such as Norvalk virus (20-30 nm) and human papilloma virus (,55 nm) diffuse through the cervical mucus as rapidly as in water, while larger (180 nm) herpes simplex virus diffuses 100-1000 times slower in the mucus compared to water [20]. In addition to being covered by mucus, on their apical side IECs have microvilli which could further hinder their contact with larger particles. We observed 40 nm NPs in the endocytic domains between the IEC microvilli, which suggested that small NPs could be internalized by an active endocytic mechanism. In cultured cells the particle size was shown to determine the extent and the pathway of their uptake. Latex particles ,200 nm are taken up via clathrin-mediated endocytosis and this uptake can be inhibited by CPZ [28,29]. As reported for cultured epithelial cells [28], we found that the internalization of 40 nm NPs by IECs can be inhibited by CPZ in vivo. However, to markedly inhibit the NP uptake at least a 5-fold higher concentration of CPZ was necessary. CPZ inhibits clathrin-mediated endocytosis [30], but can also interact with intracellular lipids and cytoskeletal regulators, thus it may affect the uptake of substrates via fluidphase macropinocytosis [31]. The use of endocytosis inhibitors is still a preferred tool for in vivo studies. However the potentially poor specificity of some inhibitors should always be considered when results of in vivo studies are interpreted. Although CPZ is not an absolutely specific inhibitor of clathrin-mediated-endocytosis [30], the use of CPZ is a good initial approach for distinguishing clathrin-mediated from other endocytic pathways.
In several experiments we observed that the amount of 40 nm NPs in the MLNs was significantly decreased (but not abolished) by CPZ, while the entry of dextran via GAPs was unaffected. The uptake of NPs by M cells in the Peyer's patches may account for the presence of NPs in the MLNs of CPZ-treated mice however, we do not have experimental data on whether CPZ affects antigen uptake by M cells. In addition, the varying presence of chyme and the mixing of the intestinal contents in vivo due to intestinal peristalsis may lead to an uneven distribution and absorption of the small volume of PBS/CPZ and NPs in various areas of the SI. Therefore, delineating the endocytic pathways by which IECs internalize NPs in vivo still remains a daunting task. Under similar experimental conditions, internalization of 20 nm NPs was unaffected by administration of CPZ or genistein, even when used at higher concentrations. These findings suggest that depending on their size, NPs may be internalized via distinct pathways, as reported for cultured cells [28]. This possibility is supported by the observation that Ova antigen conjugated to 20 nm NPs reaches the LP and the serosa of the SI in an immunologically relevant form and can be detected with anti-Ova antibodies. We do not believe that administration of scopolamine for in vivo imaging has a significant effect on NP internalization for two reasons. First, scopolamine was administered 10-15 minutes before imaging and at least 20-30 minutes after NP administration and second, no scopolamine was administered to mice that were used for quantifying the NPs in the MLNs by IFM and by spectrofluorophotometry. NPs and virus-like particles are increasingly being used as vehicles for delivery of antigens and targeted delivery of drugs [32,33]. Larger sized NPs (.200 nm) have been studied extensively due to their ability to carry larger amounts of cargo, however, there has been little work done in using NPs smaller than 100 nm for these applications. Further characterizing the uptake and trafficking of the NPs and the immune responses to NPconjugated antigens will be important for understanding how tolerance and immunity to intestinal antigens are generated. This work will also be important for the development of more effective mucosal vaccines and therapies. Movie S3 Entry of dextran into the LP of the villi via GAPs. Dextran and 20 nm NPs were administered into the SI lumen and 40 minutes later the lumen side of the intestine was imaged in vivo with a confocal microscope. Dextran (green) can be seen entering the LP via GAPs, while NPs highlight the IECs and can be seen in the LP of the villi. Movie shows 27 Z stacks taken from the tip of the villi to a depth of 70 mm.