HeLa and NIH 3T3 fibroblast cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM) (11960-044, Gibco by Life technologies) supplemented with 10% fetal bovine serum (F0926, Sigma), 20 mM L-glutamine (25030081, Gibco by Life technologies) and 10 μg/mL gentamicin (Life technologies). For subcultures, cells were trypsinised with 0.05% trypsin-EDTA solution (Life technologies). HaCaT cells were cultivated in 1 part F12 (11765-054, Gibco by Life technologies) and 3 parts DMEM supplemented with 10% fetal bovine serum FCS-SA, Labtech, 10 μg/mL gentamicin (Life technologies), 20 mM L-glutamine, Insulin (I9278, Sigma), and HCE cocktail containing hydrocortisone (AC35245-0010, Fisher Scientific), cholera toxin (C8052, Sigma) and epidermal growth factor (AF-100-15, PeproTech) (final concentrations in media were 10⁻¹⁰ M, 0.5 μg/mL and 10 ng/mL, respectively; Celis, 2006). Cells were trypsinised with 0.5% trypsin-EDTA (10x) solution (15400-054, Gibco by Life technologies) diluted to 5x with versene solution (15040-066, Gibco by Life technologies). All cell lines were cultured at 37°C with 5% CO₂. To induce HaCaT differentiation, cells were grown until confluence, and cell culture media was substituted by DMEM:F12 (Life technologies) supplemented with 1% FBS and 2 mM calcium chloride (CaCl_2; Sigma).

HeLa cells were used to propagate C. trachomatis serovar L2 (LGV-434) (C. trachomatis L2) or C. trachomatis serovar D (C. trachomatis D) (kindly provided by Dr. Grieshaber, University of Idaho). Infected cells were harvested, and C. trachomatis was purified by discontinuous density gradient centrifugation in gastrografin (Bayer). For infection, cell culture monolayers were plated to 70–80% confluence on a glass coverslip except for the calcium-differentiated experiments where HaCaT cells were plated to 100% confluence. Monolayers were infected at an MOI (multiplicity of infection) of 1 or 5 and centrifuged for 5 min at 1,000 rpm, 4°C, unless otherwise stated. Negative controls (mock-infected) were incubated with cell culture media only. Cells were incubated at 37°C with 5% CO₂ for the specific time points. Inoculum for 3D cultures infection was calculated considering a confluent monolayer, and were infected at an MOI of 5 by introducing C. trachomatis inoculum (10⁹ IFU/mL for C. trachomatis L2 and 10⁸ IFU/mL C. trachomatis D) at the top of the cultures in a total volume of 80 μL per culture. The cultures were allowed to rest for 15 min prior to returning them to the incubator for the specific time points. For penicillin experiments, 3D cultures were infected at the day of liquid-air interface. After 2 h of infection, media containing penicillin (final concentration 100 μL/mL, ~15 U/mL) was added to the bottom of the insert. For each time point of penicillin recovery, wells were washed with cell culture media twice followed by the addition of fresh cell culture media until the end of the experiment.

The experiments were conducted following the appropriate protocols commensurate with Biosafety Level 2 procedures.

Triplicates of infected monolayers were washed with Hank's Balanced Salt Solution (HBSS) (Gibco by Life technologies). Cells were lysed by addition of ice-cold water. Cell suspensions were diluted (1:50) in the adequate cell culture media and used to re-infect fresh monolayers, in triplicate. Plates were incubated for 1 h at 37°C with 5% CO₂ and the inoculum replaced with fresh cell culture media. Plates were returned to the incubator for an additional 24 h. Samples were then fixed with 4% PFA (Sigma) for 15 min at room temperature (RT) and washed three times with Phosphate Buffered Saline (PBS).

A collagen solution was prepared following the manufacturers protocol for RAFT™ reagent kit (016-0R94, Lonza) for 24-well plates. When indicated, NIH 3T3 fibroblasts were added to the collagen solution and pipetted to a 24-well culture plate. Plates were incubated at 37°C for 15 min. After incubation, absorbers were applied on top of each collagen gel for 15 min. Then, the absorbers were removed and cell culture medium containing 3 × 10⁵ cells/mL of HaCaT cells were added. Plates were incubated at 37°C with 5% CO₂ in submerged cell culture media until cells reached confluence. Cell-seeded collagen gels were transferred to 6-well plate cell culture inserts (PICM03050, Millicell®) at the air-liquid interface (non-submerged) for the indicated times. Media was changed every other day. On the specified time points, the cell culture wells were washed three times with HBSS and OCT-embedded (23-730-571, Fisher Scientific) followed by rapid freeze in isopentane chilled with liquid nitrogen. There were variability in the raft culture thickness, with the majority of the observed differences ranging from 1 to 2 cell layers, but minimal variability in differentiation. One possible cause is how many times the HaCaT cells were passaged. We did not use HaCaT cells passaged beyond 15 passages.

Cryo-embedded sections were washed with distillated water for 5 min. Nuclei were stained with haematoxylin for 5 min, washed in tap water, differentiated by immersing in 1% acid alcohol, and rinsed in water. The cytoplasm was stained with eosin for 5 min. Sections were dehydrated by ascending grades of ethanol followed by xylene. Sections were mounted with Immu-Mount mounting medium (Thermo Scientific). Brightfield pictures were obtained on the Leica Camera w/Leitz & Fluorescence Stereo Microscope. Immunofluorescence was performed on frozen and OCT embedded sections. Sections were air-dried and then fixed with 4% PFA for 15 min. Cells grown on glass coverslips were rinsed three times with PBS and fixed with 4% PFA for 15 min at RT. Fixed samples were rinsed with PBS and permeabilized with 0.2% Triton X-100 in PBS for 5 min at RT with agitation. Samples were washed three times with PBS, and blocked with 5% BSA for 30 min, at RT. Then, sections or coverslips were incubated with anti-K14 (ab51054, Abcam), anti-K10 (ab76318, Abcam), anti-Chlamydia LPS (ab21180 or ab62708 from Abcam or NB100-62449 from Novus), anti-Chlamydia (PA1-25652 from Thermo Fisher Scientific) anti-filaggrin (ab3137, Abcam) or (ab92742, Abcam). Phalloidin (A12379, Thermo Scientific) and DAPI were used to stain F-actin and DNA, respectively. Samples were then rinsed 3 times with PBS and incubated with appropriate Alexa Fluor conjugated secondary antibodies (Thermo Fisher Scientific). Coverslips and sections were mounted onto microscope glass slides using Immu-Mount mounting medium (Thermo Scientific). All samples visualization was performed on a laser scanning confocal microscope (Leica SP8 Point Scanning Confocal or Leica SP8-X White Light Point Scanning Confocal). Images were processed using ImageJ.

3D cultures were washed twice with HBSS. Cultures were fixed with 2.5% glutaraldehyde in Phosphate Buffered Saline (PBS) for 1 h at 4°C. Fixed cultures were washed 3 times with PBS for 5 min each and post-fixed with 1% osmium tetroxide in PBS for 30 min. Then, samples were washed 3 times with distilled water for 10 min to remove the excess of phosphate. Cultures were dehydrated in graded ethanol starting with 50, 70, 95, and 100% for 30 min each followed by infiltration with ethanol/Spurr's (1:1) solution. After 1 h, a fresh infiltration solution of ethanol/Spurr's (1:10) was added for an additional 60 min. 3D cultures were embedded in fresh Spurr's resin, placed in a 60°C oven, and polymerized overnight. Seventy to ninety nanometers of thick sections were cut on Leica UC6 or Reichert ultramicrotome. Samples were visualized on a JEOL 1400 plus + AMT UltraVUE camera and FEI TEM T20.

HaCaT cells were collected at the time of differentiation induction as a time-point zero, and then at 2, 3, 5, or 7 days post-differentiation. For each time-point, cells were washed in PBS and then lysed with RIPA buffer (AAJ60780EQE, Fisher Scientific) supplemented with protease inhibitor (Roche Diagnostics) and phosphatase inhibitor (Roche Diagnostics), following manufacturer's instructions. Samples were incubated at 4°C for 30 min followed by centrifugation at 14,000 × g for 20 min at 4°C. After supernatants were collected, protein concentrations were determined using Bradford's assay. Samples were diluted in 5x SDS sample buffer and boiled for 10 min at 100°C. Samples were resolved in a SDS 10% polyacrylamide gels and transferred to a nitrocellulose membrane (Whatman) for 60 min at 100 V by the standard wet transfer method. Immunoblotting was performed using primary antibodies against cytokeratin 10 (K10) and cytokeratin 14 (K14) (ab76318 and ab51054, respectively from Abcam), followed by the secondary antibodies polyclonal goat anti-rabbit and polyclonal rabbit anti-mouse conjugated to HRP (DAKO). Tubulin antibody (ab21058, Abcam) was used as a loading control.

HaCaT cells were plated onto the surface of collagen gels containing embedded NIH 3T3 fibroblasts, and cultured at the liquid-air interface. Previous studies showed that NIH 3T3 fibroblasts employed as feeder cells supported tissue architecture and terminal differentiation in a similar way as human dermal fibroblasts (Smola et al., 1993; Maas-Szabowski et al., 2001). Different concentrations of fibroblasts were tested to determine the optimum number required to achieve multi-layered and differentiated epithelia. Cultures with 0, 2 × 10⁴, or 4 × 10⁴ fibroblasts embedded in collagen were grown and differentiated in air-exposed cultures for 14 days (Supplementary Figure 1). Raft cultures were fixed and processed for histochemical staining. Samples cultured in the absence of fibroblasts exhibited less stratification as compared to those co-cultured with higher number of fibroblasts (Supplementary Figure 1). We also determined that in our hands, full stratification, which was defined as the formation of several layers of squamous cell in the uppermost region at 14 days of culture, required at least 4 × 10⁴ fibroblasts embedded in the 100-μm collagen matrix (Supplementary Figures 1, 2). Enucleated cells were not common, and keratinization was incomplete. These are typical features of non-keratinizing epithelia (Chateau and Boehm, 1996; Kikuchi et al., 1997; Anderson et al., 2014).

To investigate the differentiation of HaCaT cells when raised to the air-liquid interface, the expression of specific keratins was examined. Cryo-embedded sections were analyzed by immunofluorescence for specific differentiation markers, including keratin 14 (K14), keratin 10 (K10), and filaggrin. We also monitored the expression of the nuclear protein ki67, which indicated cell proliferation (Supplementary Figure 2). Cells in the basal and suprabasal layers were positive for K14. Ki67 signal was predominantly associated with the basal layer. K10 staining demonstrated some overlap with that of K14, indicating cells transitioning between differentiation states. As expected, ki67 signal did not overlap with K10. Filaggrin could be observed in the upper layers, with partial overlap with K10. We also observed well-formed desmosomes by transmission electron microscopy (Supplementary Figure 2). In summary, our 3D cultures showed the following markers for the different layers: ki67, K14, K10, and filaggrin, with some layers exhibiting overlaps of each marker tested. Together with the well-formed desmosomes, these observations indicate that our in vitro-generated stratified epithelia mimic important differentiation and stratification features of squamous epithelia.

Previous work showed that HaCaT cells were able to support chlamydial growth (Joubert and Sturm, 2011). We confirmed this finding by analyzing the production of inclusion forming units (IFU), and inclusion morphology. Both HaCaT and HeLa cells were infected with C. trachomatis L2 at an MOI of 1. Inclusion development was monitored at 2, 20, and 36 h post-infection (h p.i.), and visualized by indirect immunofluorescence using anti-chlamydial LPS antibody, followed by confocal microscopy (Supplementary Figure 3A). Inclusions within HeLa and HaCaT cells were very similar in appearance. We then determined if both cell lines could equally support the completion of the developmental cycle, i.e., production of infectious particles (EBs) (Supplementary Figure 3B). Infectious particles were harvested at different time points by hypotonic lysis, and the recovered infectious particles were added to a fresh monolayer of HeLa cells to determine harvest titers. Infection was allowed to proceed for 24 h, and samples were processed for immunofluorescence to visualize and enumerate inclusions. Both HeLa and HaCaT cells produced similar inclusion forming units (IFU) (10⁷ IFU/mL) at the latest time point monitored (36 h p.i.; Supplementary Figure 3B). The results indicated that in a monolayer configuration, HaCaT cells equally supported chlamydial development and replication as HeLa cells.

The stratified squamous tissue has a more complex epithelial organization given its various stratification and differentiation states. Three-dimensional organotypic stratified epithelial cultures (3D cultures) of HaCaT were infected, and C. trachomatis growth assessed. Air-exposed cultures were allowed to stratify and differentiate for 14 days prior to infection. On the 14th day, the inoculum was directly applied to the top of the stratified epithelia where terminally differentiated cells were located (Supplementary Figure 2). The inoculum was calculated by considering a confluent monolayer at the top of the 3D cultures, and 80 μL was introduced on top of each 3D culture. Infection was allowed to proceed for 1 or 5 days for C. trachomatis L2 (Figure 1A) and 1 or 3 days for C. trachomatis D (Supplementary Figure 4) before sample processing and analysis. Cryosections were stained for C. trachomatis using an anti-Chlamydia LPS antibody. C. trachomatis L2 inclusions were only observed in the top layers of the stratified epithelia with no apparent dissemination to the lower layers for both 1 and 5 days post-infection (d p.i.) samples (Figure 1B). Similarly, C. trachomatis D infection showed no apparent dissemination, and in contrast to C. trachomatis L2, inclusions were almost absent by immunofluorescent analysis by 1 d p.i. (Supplementary Figure 4). By 3 d p.i. no inclusions were detected for C. trachomatis D, which precluded analysis of developmental forms by transmission electron microscopy (TEM). We evaluated the morphology of C. trachomatis L2 developmental forms at 1, 3, and 5 d p.i. (Figure 1C) from three independent experiments. The majority of the inclusions contained small numbers of reticulate bodies (RBs) in both 1 and 3 days post-infection samples. We also observed very few infected cells in the top layer in the 3 day samples, which could be attributed to the natural shedding occurring in the uppermost layer of the epithelium (Figure 1C, black arrowheads). By 5 days post-infection no inclusions were found by TEM. We examined the morphology of C. trachomatis L2 forms within the inclusions, and quantified the area for each form. We measured the area rather than the diameter given the irregular shape of some of the RBs. RBs that appeared to be undergoing binary fission were excluded. TEM images across three independent experiments were obtained for 1 and 3 d p.i. when possible. Previous studies have established that the area of an RB is between 0.4 and 0.8 μm², and the area of an EB is expected to be between 0.07 and 0.12 μm² (Wyrick, 2010). Indeed, from cells containing newly formed EBs, such as those shown in Figure 3, we found our results to be consistent with previous reports (Figure 1D). The inclusions found in the uppermost layers of the 3D epithelium contained Chlamydiae with areas between 0.25 and 0.65 μm² (Figure 1D), which are consistent with the area of an intermediate body (IB), although we could not exclude the possibility that some of the organisms categorized as IBs may instead be RBs sectioned above or below the “equator.” Overall, when compared against the growth kinetics obtained for HeLa and HaCaT cells in the monolayer configuration, C. trachomatis development was severely delayed in the stratified and differentiated HaCaT cells. Despite allowing for the infection to progress for up to 5 days, the inclusions remained morphologically immature, with very small numbers of Chlamydiae residing within the cells, and the absence of late developmental forms, e.g., elementary bodies. This was in contrast to inclusions in HaCaT monolayers, where the majority was readily visible with multiple organisms after 1 day post-infection (Supplementary Figure 3).

It is possible that the observed delay in chlamydial growth in stratified epithelia was due to the differentiation state of the cells. To investigate this, we took advantage of a known property of HaCaT cells—the induction of differentiation by the addition of high calcium concentration to the culture medium (Hennings et al., 1980; Deyrieux and Wilson, 2007). After seeding on coverslips and growing to full confluence, media containing 1% FBS supplemented with 2 mM Ca²⁺ was added to the cells to induce differentiation. Differentiation was confirmed independently by monitoring the changes in expression of the keratinocyte differentiation markers K10 and K14 by fluorescence microscopy and Western blot. In samples exposed to high [Ca²⁺], we observed decreased expression of K14 along with the concomitant increase in K10 expression. In the untreated controls, cells maintained the expression of K14, and remained negative for K10 (Figures 2A,B). In a parallel experiment, Ca²⁺-differentiated HaCaT cells were infected with C. trachomatis serovar L2 by inoculating the monolayer without centrifugation (Figure 2C). C. trachomatis L2 were introduced to cultures and incubated for 1 h at 37°C followed by a washing step to removed unadhered EBs. Fresh media was added, and the cells incubated for 24 h. As shown in Figure 2C, serovar L2 was able to infect and form inclusions regardless of the duration of differentiation. Inclusion sizes were heterogeneous, which is likely due to the asynchronous method of infection, e.g., without centrifugation. To better assess chlamydial development, we synchronously inoculated with serovar L2 samples of HaCaT monolayers at different times of Ca²⁺-induced differentiation, including an undifferentiated control, and monitored IFU production at 24 h p.i. HeLa cells were used as control. All treatment groups yielded similar levels of IFU (Figure 2D). For this reason, we concluded that Ca²⁺-induced differentiation of HaCaT cells had no effect on C. trachomatis serovar L2 growth and development.

We investigated the interaction of C. trachomatis serovar L2 with the undifferentiated (i.e., ki67-positive layer) and differentiating layers of the 3D cultures (i.e., K10-positive layers). In vivo, these cells become accessible to Chlamydia through microabrasions that occur naturally in the stratified squamous epithelium after consensual intercourse (Norvell et al., 1984). Pathogens such as HPV can take advantage of the microabrasions in the tissue to reach more hospitable cells where they can establish infection. Therefore, in our system, we investigated if the same basal cell types are much better hosts for C. trachomatis than the fully differentiated cells found in the top layer. We attempted to mimic microabrations in our raft cultures by puncturing the stratified epithelia with a small gauge needle. However, we encountered two major technical difficulties. First, the depth and angle of the puncture could not be reliably reproduced because of the unevenness of the surface of the raft cultures. Second, consistently capturing the entire depth of the punctured areas during histological sectioning proved to be difficult. Therefore, we resorted to the direct inoculation of undifferentiated or differentiating layers. While we were able to perform infections of layers at a defined differentiation state, we are cognizant of an important caveat to our strategy, which is that infection was performed in the absence of the additional layers that surrounds the infection site, and thus could potentially affect its progression. Inclusion growth and C. trachomatis L2 developmental forms were monitored during infection of undifferentiated layers (i.e., prior to lifting at the air-liquid interface) and intermediately differentiating layers when the culture was lifted at the air-liquid interface for a defined duration prior to C. trachomatis L2 infection (Figures 3A, 4A). When undifferentiated cells were infected with C. trachomatis, and then allowed to stratify and differentiate, inclusions were present at 1 d p.i. At 3 d p.i. (and thus 3 d post-lifting to air-liquid interface) mature inclusions with clearly visible EBs and RBs (Figure 3B) were observed. C. trachomatis L2 developmental forms present at 1 and 3 d p.i. were characterized (Figure 3C). We acquired TEM images, and measured the area of the organisms from 1- to 3-d samples. In the two inclusions found at 1 d p.i., the area of the C. trachomatis L2 organisms was between 0.2 and 0.3 μm², which was smaller than RBs (0.4–0.65 μm²) from samples shown in Figure 1D (dashed line) and Figure 3C (yellow line). The lack of condensed nucleoid indicated that these forms may have been in the middle of the EB-to-RB transition. However, the size distribution of chlamydial organisms from five inclusions at 3 d p.i. indicated the presence of EBs, IBs, and RBs, suggesting that between 1 and 3 d p.i., C. trachomatis L2 replicated and completed its developmental cycle. The majority of the C. trachomatis L2 forms had areas of 0.15 μm². The areas of RBs ranged from 0.4 to 0.65 μm², although a few large ones were found (0.95–1.15 μm²). Based on morphometric measurements, we conclude that proliferating undifferentiated HaCaT cells supported the completion of the developmental cycle, with inclusions harboring all known developmental forms.

We also investigated chlamydial growth when infecting layers in the middle of differentiation (i.e., positive for K10) by inoculating the raft cultures at day 9 post-lifting at the air-liquid interface (Figure 4A). At this time point, it was clear that there were less layers when compared to a fully stratified day 19 raft culture (Supplementary Figure 2), and the upper layers expressed K10 (Figure 4B). We focused our attention on the 3 and 5 d p.i. groups as they yielded sufficient numbers of inclusions for analysis (Figure 4B). Quantification of C. trachomatis L2 particle area from two independent experiments showed that in both time points (3 and 5 d p.i.) the majority of particles harbored condensed nucleoid, and had areas between 0.05 and 0.15 μm² consistent with EB forms (Figure 4C). RBs were found in the 3 d p.i. samples ranging from 0.35 to 0.85 μm², and one enlarged RB with 0.95 μm². By 5 d p.i. the trend was skewed toward EBs with some intermediate forms with areas ranging from 0.25 to 0.35 μm², and RBs ranging between 0.45 and 0.75 μm² (Figure 4C), indicating ongoing development. These observations were in stark contrast to those samples where the top-most layers had inclusions with condensed nucleoid-negative forms that resembled neither EBs nor RBs with regards to size (Figure 1C).

We also evaluated the effects of differentiation on infection by C. trachomatis serovar D. Being a genital strain, we wanted to determine if it exhibited some form of adaptation for better growth in differentiating cells. As shown above, serovar D, like serovar L2 was able to grow in the basal layer, but not in the terminally differentiated, filaggrin-positive layers. Here, we asked if serovar D, would grow and replicate in differentiating cells in two independent configurations. Differentiation in HaCaT monolayers or stratified raft cultures was induced by exposure for different duration to high [Ca²⁺] at or lifting at the air-liquid interface, respectively. Differentiation was followed by inoculation in Ca²⁺-free media for 1 h without centrifugation. For the raft culture, infection was performed at 9 d post-differentiation, which is well ahead of the 19 days it took to acquire a fully differentiated, e.g., filaggrin-positive, upper layers. As shown in Figure 5A, serovar D infection decreased noticeably in samples exposed to high [Ca²⁺] for 5 and 7 days (data from two independent experiments). Very few, if any inclusions were present in these samples. Infection of the intermediately differentiated raft cultures also exhibited very few infected cells and greatly diminished levels of dissemination, with C. trachomatis serovar D organisms only found at the upper layers (Figure 5B). We conclude that at least for serovar D, the differentiation state of the host cell determined how well this serovar grew.

To gain initial insight into how infection might disseminate in stratified epithelia, HaCaT cells at 0 d post-lifting at the air-liquid interface were infected with C. trachomatis L2, and immediately raised to initiate differentiation. Infection was also allowed to proceed for shorter periods of time of 5 days for C. trachomatis L2, and 1 and 3 days for C. trachomatis D (Supplementary Figure 5A). At this time point it was evident that the 3D cultures were initiating the stratification process and only 1–2 cell layers have formed, and that the basal layer expressed the proliferative marker ki67 (Supplementary Figure 5B). C. trachomatis L2 and D infections were detectable in cryosections analyzed from four 3D cultures for C. trachomatis D and two independent experiments for C. trachomatis L2 (Supplementary Figure 5C). We focused on C. trachomatis L2 samples to determine if dissemination could be inhibited by penicillin. Infection was allowed to proceed for 19 days in the absence or presence of the antibiotic (Figure 6A). The samples were processed for immunostaining for C. trachomatis to reveal the extent of dissemination of infection. In the absence of penicillin for 19 d, infection disseminated throughout the 3D culture (Figures 6B,C), which was expected for a productively infected epithelium. Similar results were observed for C. trachomatis D-infected 3D cultures (Supplementary Figures 5D,E). Addition of penicillin significantly limited C. trachomatis L2 infection, which were confined to the upper layers. The upper layer location is likely from the continued upward displacement of cells harboring penicillin-arrested C. trachomatis in the stratifying raft culture.

We also included parallel cultures where penicillin was removed at either 5 or 10 d p.i. to determine if disseminated infection could be restored for the remainder of the experiment. These groups were left untreated for 14 or 9 d, respectively. As shown in Figure 6D, the extent of dissemination did not appreciably differ from the samples that were continuously exposed to the antibiotic. Chlamydia LPS was detected at the upper layers only. The lack of restored growth could be due to the continued presence of residual penicillin in the upper layers despite removal of penicillin.

Based on the results, C. trachomatis L2 appeared to exhibit preference for the undifferentiated or differentiating layers over terminally differentiated ones, while serovar D was negatively influenced by the differentiation state of the host cell. Infection of the filaggrin-positive differentiated cells in the uppermost layer was unproductive, while infection of the proliferating basal cells yielded inclusions containing EB-like organisms, indicating the completion of the developmental cycle. When allowed to proceed for 19 days, infection disseminated throughout the stratified epithelium, resulting in the significant destruction of the 3D culture. The extensive dissemination of infection was consistent with the apparent completion of the developmental cycle in the ki67-positive undifferentiated basal cell layer (Figure 3B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.