Cells cultured on solid surfaces are exposed to an environment that is different from the one surrounding cells in a living multicellular organism. In order to determine whether environment and spatial organization affect the architecture of subcellular compartments, we compared cells grown either on glass or in a 3D matrix with cells in intact tissues. Specifically, we used a cell line derived from human submandibular glands (HSG) (Figure 1A upper panel) [29,30] which forms polarized acinar-like structures that morphologically resemble the acini in intact SGs, when cultured in matrigel (Figure 1A center panel) [31,32]. First, HSG cells were stained with phalloidin, to reveal the organization of the actin cytoskeleton (Figure 1B). In HSG grown either on glass or plastic surfaces actin formed stress fibers and the nuclei where localized in the center of the cells (Figure 1B upper panel). When grown in matrigel, cells formed polarized acini with the nuclei located at the basolateral plasma membrane and actin reorganized in thick bundles at the apical cell cortex (Figure 1B middle panel, arrow). Notably, in intact SGs the organization of actin was more complex. Indeed, actin decorated very narrow canaliculi that are localized exclusively at the apical plasma membrane (Figure 1B lower panels, arrows). These results suggest that in SGs-derived cells removing the constrains imposed by the interaction with a solid substrate and establishing cell-cell interactions in three dimensions is sufficient to promote a dramatic rearrangement of the architecture of the cells, the actin cytoskeleton, and to initiate polarity. Similar conclusions were derived for mouse fibroblasts grown under similar conditions and compared with mouse tissue [33]. Next, we analyzed the localization of transferrin receptor (TfnR), a marker for early and recycling endosomes [34], and VAMP8, a SNARE protein that has been implicated in regulated exocytosis in several systems [35,36]. Interestingly, in both 2D and 3D cell cultures, VAMP8 colocalized with TfnR (Figure 1C upper and center panels), whereas in the intact tissue they did not. More specifically, in SGs, VAMP8 localized in the subapical area of the acinar cells, whereas TfnR localized in intracellular vesicles closer to the perinuclear area (Figure 1C lower panels). These findings strongly suggest that both the spatial organization and the unique combination of molecules that are present in the intact tissue substantially affect the organization of the intracellular compartments.

These morphological differences raise a fundamental question on whether the dynamics of membrane trafficking events may vary between cultured cells and live animals. Since we have recently shown that both the regulation and the modality of exocytosis in live animals are significantly different from what reported in in vitro model systems [20], here we sought to extend our investigation to endocytic events. We imaged fluorescently-labeled Tfn and dextran in the submandibular SGs of live rats, two molecules that are internalized by clathrin-dependent and clathrin-independent pathways, respectively [37,38]. To this aim, Texas-Red Tfn (TXR-Tfn) and Alexa-488 dextran (488-D) were injected systemically in the tail artery of anesthetized rats, as previously described [17,26]. Since excess of Tfn can be diverted to non-clathrin pathways, we injected 200 μg of fluorescent Tfn an amount equivalent to approximately 1:100 of the Tfn plasma levels, as estimated by others [39]. Moreover, in order to minimize any difference between the diffusion of the two probes either from the vasculature or throughout the stroma, we used a 70-kDa dextran that has a MW close to that of Tfn (80 kDa). We found that both molecules appeared in the vasculature and diffused into the stroma a few seconds after injection (Figure 2A, arrow). Three min after the injection only dextran was internalized in stromal cells (Figure 2A arrowheads) whereas Tfn was detected in endosomal vesicles after 10 min (not shown). Interestingly, endocytosis was evident mainly in stromal cells while the polarized epithelium did not seem to internalize the fluorescent Tfn and dextran. Furthermore, the observed Tfn internalization by stromal cells was slower than what has been reported in tissue culture. One explanation for the delayed internalization of Tfn with respect to dextran is a difference in the diffusion rate out of the vasculature, in spite of their similar MW. To overcome this potential issue we delivered the probes by a direct injection into the exposed glands, and imaged in an area distant from the injection site (Figure 2B). Interestingly, although both molecules were clearly observed in the stroma (Figure 2B left panels, arrows), only dextran was internalized by stromal cells within 5 min from the injection (Figure 2B left panels, arrowhead). After 20 min, Tfn was internalized and localized in endosomes different than those containing dextran (Figure 2B, center panels, asterisks). After 30 min, both molecules overlapped in a small subpopulation of endosomes (Figure 2B, right panels). Notably, the slower rate of Tfn internalization was not due to the lack of expression of TfnR and/or clathrin, as shown in Figure 2C, suggesting that in vivo Tfn uptake may be regulated differently than in vitro.

Indeed, in cell culture Tfn has been reported to be internalized in few seconds, whereas dextran exhibits a slower kinetics of internalization [40]. To determine whether the kinetics of internalization observed in live animals is a unique property of the SGs stromal cells, we developed a procedure to isolate and culture them on glass. Specifically, a 70 kDa Cy5-dextran was injected in the SGs, and after one h they were excised, minced, and the stromal cells were enzymatically isolated, as described in the experimental procedures. Cells were then seeded on glass coverslips and the presence of dextran was assessed by confocal microscopy (Figure 3A). After 24 h, adherent cells were incubated with fluorescent 488-D and TXR-Tfn. Strikingly, Tfn was internalized within 30 seconds, whereas dextran was internalized after 10–15 min (Figure 3B). After 30 min, both probes were detected in intracellular structures and no colocalization was detected.

These results suggest that endocytic pathways undergo substantial modifications when cells are removed from their in vivo environment. These modifications can be interpreted as an adaptive response to the new environment. For example, we can speculate that the increase in the kinetics of Tfn internalization may be related to a need to increase iron metabolism, as protection from either oxidative stress or bacterial infection. Indeed, the increase in oxygen levels and the lack of cytokines from the immune system may work as a trigger to alter these processes. Alternatively, it is possible that the interactions between cells and extracellular matrix in vivo may negatively regulate receptor-mediated endocytosis. Understanding the mechanisms underlying these modifications may provide novel information on the regulation of endocytosis under physiological conditions.

All the fluorescent molecules and the secondary antibodies were purchased from Invitrogen (Carlsbad, CA, USA). Mouse anti-clathrin and rabbit anti VAMP8 antibodies were from Abcam Inc. (Cambridge, MA, USA), mouse anti-TfnR antibody was from Zymed (Carlsbad, CA, USA).

Sprague–Dawley male rats weighing 150–250 g were obtained from Harlan Laboratories Inc. (Frederick, MD, USA). The animals were acclimated for one week before used for the procedures. Water and food were provided ad libitum. All the experiments were approved by the National Institute of Dental and Craniofacial Research (NIDCR, National Institute of Health, Bethesda, MD, USA) Animal Care and Use Committee. The animals were anesthetized by an IM injection of a mixture of Ketamine and Xylazine (100 mg/kg an 20 mg/kg respectively) with additional injections as needed. Intravital microscopy and injection of the probes were preformed as described in [17,24]

An IX81 inverted confocal microscope (Olympus, Melville, NY, USA) was modified to perform two-photon microscopy as described previously [26]. The excised glands and cultured cells were imaged in the inverted setting while time lapse imaging on the live animal was performed in the upright configuration using an objective inverter (LSM Technology Inc., Shrewsbury, PA, USA) [41]. For the time lapse imaging the acquisition speed was set to 0.3 frames/sec. All the images and movies were acquired using a UPLSAPO 60X NA 1.2 water immersion objective (Olympus).

When needed, the background noise was reduced by applying to each image one or two rounds of a 2 × 2 pixel low-pass filter by using METAMORPH (Molecular Devices). Brightness, contrast and gamma correction were applied. Volume rendering was performed using IMARIS 5.6 and 6.0 64 bit (Bitplane). The final preparation of the images was conducted with ADOBE PHOTOSHOP CS.

HSG cells [30] were cultured in Dulbecco’s modified Eagle’s medium/Ham’s F-12 (1:1) (Invitrogen, CA) in the presence of 5% fetal bovine serum, 100 U/mL penicillin, and 100 mg/mL streptomycin (Sigma-Aldrich, MO) at 37 °C in a 5% CO₂ humidified atmosphere.

Growth factor-reduced matrigel (Trevigen, MD, USA) was diluted in complete medium at the appropriate concentration (2.0–6.0 mg/mL), added to a 12-well tissue culture dish, and incubated for 1 h at 37 °C before culturing of the cells. 50 cells/mm² were added onto the pre-solidified matrigel and grown for different times. Acinar cells were recovered from the matrigel by incubating in cell recovery solution (Trevigen, MD) for 30 min on ice. Acinar cells released from the matrigel were spread on glass coverslips and dehydrated at 37 °C for 15 min. Cells were then fixed in 2% formaldehyde/PBS, incubated first in blocking solution (10% FBS, 0.02% sodium azide in PBS) for 15 min, and later with various primary antibodies for 1 h at RT. Cells were then washed and incubated with the appropriate secondary antibodies for 1 h at RT. After three washes with blocking solution, nuclei were labeled by incubating the cells with Hoechst 33342 before mounting.

Isolated glands were fixed for 20 min in 0.05% Glutharaldehyde/4% formaldehyde in 0.2M Hepes pH 7.3 and then transferred for 1 h in 4% formaldehyde. The glands were sliced from the surface and fixed again in 4% formaldehyde for 30 min. The fixative was washed in PBS and the slices were blocked overnight at 4 °C in PBS, 10% fetal bovine serum, 0.04% saponin (blocking solution). The slices were incubated for 1 h with the primary antibody in blocking solution, washed 3 × 10’ in PBS and then incubated with the appropriate secondary antibody for 1 h at R.T. The samples were washed 3 × 10’ in PBS and then imaged by either two-photon or confocal microscopy.

Sprague-Dawley rats were anesthetized, the submandibular SGs exposed, and 50 μL of 10 μg/mL Cy5 dextran were injected on the side of each gland. After one h, the glands were excised, placed in ice-cold 1% BSA in DMEM. The glands were then injected with the same medium using a syringe equipped with a 30 GG needle. After 5-10 injections the connective tissue was separated from the SGs lobules, collected, and incubated for 1 h at 37 °C in 2 mg/mL collagenase and 0.125 mg/mL dispase (Invitrogen). The tissue was homogenized by continuous pipetting with a 10 mL pipette and filtered through a sterile gauze to remove large fragments. The cells suspension was centrifuged at 2,000 rpm for 5 min and the pellet suspended in fresh DMEM containing 5% fetal bovine serum, 100 U/mL penicillin, and 100 mg/mL streptomycin (Sigma-Aldrich, MO, USA). The cells suspension was seeded on sterilized glass coverslips and grown at 37 °C in a 5% CO₂ humidified atmosphere. After 24 h the coverslips were incubated for 30 min with 0.5% BSA in DMEM and the internalization of Tfn and dextran was performed in the same medium.