E6/E7 Functional Differences among Two Natural Human Papillomavirus 18 Variants in Human Keratinocytes

It is suggested that HPV-18 variants from the A lineage have higher oncogenic potential compared to B variants. Some studies show uneven distribution of HPV-18 variants in cervical adenocarcinomas and squamous cell carcinomas. Regarding HPV-18 variants’ functions, the few studies reported focus on E6, and none were performed using natural host cells. Here, we immortalized primary human keratinocytes (PHKs) with E6/E7 of HPV-18 A1 and B1 sublineages and functionally characterized these cells. PHK18A1 reached immortalization significantly faster than PHK18B1 and formed a higher number of colonies in monolayer and 3D cultures. Moreover, PHK18A1 showed greater invasion ability and higher resistance to apoptosis induced by actinomycin-D. Nevertheless, no differences were observed regarding morphology, proliferation after immortalization, migration, or epithelial development in raft cultures. Noteworthy, our study highlights qualitative differences among HPV-18 A1 and B1 immortalized PHKs: in contrast to PHK18A1, which formed more compact colonies and spheroids of firmly grouped cells and tended to invade and migrate as clustered cells, morphologically, PHK18B1 colonies and spheroids were looser, and migration and invasion of single cells were observed. Although these observations may be relevant for the association of these variants with cervical cancer of different histological subtypes, further studies are warranted to elucidate the mechanisms behind these findings.


Introduction
Human papillomavirus (HPV) responds for virtually all cases of cervical cancer (CC). Worldwide, cervical squamous cell carcinomas (SCCs) are primarily associated with HPV-16 infection (60%), while HPV-16 and -18 are similarly prevalent in adenocarcinomas (ADCs) [1]. Based on a fragment of the long control region (LCR), HPV-18 genetic variants were initially classified into three branches of phylogenetic and geographical relatedness: African (Af), Amerindian (As + AI or American Indian or East Asian), and European (E) [2]. More recently, the nomenclature of HPV-18 variants was revised, and these are now categorized into variant lineages (1-2% sequence difference; designated by letters) and sublineages (0.5-1% sequence difference; designated by numbers) based on whole-genome sequencing. Thus, now while A1/A2 and A3/A4 HPV-18 sublineages correspond to previously nominated As + AI and E variants, respectively, the HPV-18 B lineage comprises previously nominated Af variants [3,4].

Spheroid Formation Assay
Cells were seeded at 2 × 10 4 in six-wells low-attachment plates (Corning Inc., Corning, New York, NY, USA) and maintained in culture for 14 days. Cells were also seeded at 5 × 10 3 in 0.5% soft agar (Invitrogen) in 24-well plates. After 15 days, colonies were stained with 1 mg/mL MTT (Sigma-Aldrich). Finally, cells were grown until 80% confluence and incubated with Nanoshuttle-PL magnetic nanoparticles (Greiner Bio-One GmbH, Frickenhausen, Germany) for 16 h. Cells were then seeded at 1 × 10 4 cells in a 96-well low-attachment plate (Greiner Bio-One GmbH, Frickenhausen, Germany), and the spheroid drive magnetic plate was placed underneath for 24 h to induce spheroid formation, which was further monitored for 15 days.

Cell Migration and Invasion
Cells were plated at 5 × 10 5 in 12-well plates. After 24 h, cells were incubated with 10 µM mitomycin C (Sigma-Aldrich) and, scrapes were performed throughout wells. To ensure cells stopped proliferating due to mitomycin C, cells were fixed, incubated with propidium iodide, and analyzed by flow cytometry. The invasion was evaluated using the QCM High sensitivity non-cross-linked collagen invasion assay (Millipore, Darmstadt, Germany) following the manufacturer's guidelines. Cells were seeded at 2.5 × 10 5 within upper chambers in nonsupplemented KSFM and as chemoattractant KSFM 15% FBS was added to lower chambers. After 72 h, inserts were stained and immersed in an extraction medium, and absorbance measurement at 540 nm was performed.

Raft Cultures and Immunohistochemistry
Immortalized PHK18A1 and PHK18B1, and parental PHKs at p0 were submitted to epithelial raft cultures, as described [28]. Briefly, parental PHK, PHK18A1, and PHK18B1 were seeded on top dermal equivalents (2 × 10 5 cells/equivalent) composed of rat tail type 1 collagen (Corning Inc., Corning, NY, USA) and 3T3-J2 fibroblasts. After 24 h, the rafts were transferred to the medium-air interface and maintained for 9 days to allow cell growth and tissue stratification. We performed two independent raft experiments with at least six replicates. Epitheliums were fixed in formaldehyde 2%, paraffin-embedded, and tissue sections obtained for histological analysis or immunohistochemistry (IHQ). Sections were probed for cytokeratin 10 (CK10) (ab76318, abcam), p16 (sc-56330, Santa Cruz, Dallas, TX, USA), and PCNA (ab29, abcam) using the Novolink Max Polymer Detection System (Leica Biosystems, Wetzlar, Germany). CC and normal epithelia were used as positive control samples for p16 and CK10 immunostaining, respectively.

Statistical Analysis
All analyses were conducted using the Statistix 8 program for Windows (Analytical Software, Tallahassee, FL, USA). The T-test was used to compare results obtained for nontransduced PHKs to those of PHK18A1 and PHK18B1. Significance levels were set at 0.05.

Immortalized PHK18A1 and PHK18B1 Show No Difference in Proliferation
Since no differences were observed between the two batches of transduced PHKs with the same HPV-18 lineage variant concerning the time required to reach p30, doubling time, and cell morphology, we decided to follow our analysis using solely transduced PHK pool 1. The proliferation rate of immortalized PHKs (p30) was assessed using three distinct approaches as follows. For both immortalized PHKs, we observed an increase in PCNA levels along passages, which was, however, more evident for PHK18B1 ( Figure 1D). Furthermore, despite we observed that both immortalized PHKs proliferated significantly faster than parental cells (PHK18A1 versus PHKs: p = 0.00; PHK18B1 versus PHKs: p = 0.01), proliferation rates were not significantly different between variants (p = 0.38) ( Figure 1E). Finally, we observed that in line with Figure 1E, Ki-67 protein levels also increased in cells over time in culture, but no significant difference was observed between variants after 6 days in culture (p = 0.19) ( Figure 1F). Taken together, our data show that while PHK18A1 reached immortalization at an earlier time point, compared to PHK18B1, after immortalization, these cells proliferated at a similar ratio.
3.3. PHK18A1 and PHK18B1 Exhibit Similar E6*I, E6, E7, p53, p16, pRb, and p21 levels The reason why the E6*I transcript is produced in HR-HPVs infected cells is not completely understood. However, it has been shown that the alternative processing that results in increased E6*I levels favors E7 expression [29]. Among HR viral types, HPV-18 infected cells exhibited the highest levels of E6*I [30]. Here, we observed similar levels of E6 + E6*I, E6, and E7 transcripts in cells transduced with HPV-18 E6/E7, regardless of HPV-18 sublineage variant or culture passage (Figure 2A).
Since the correlation between the amount of a transcript and the corresponding protein is not always direct, E6 protein levels were further evaluated. We observed that PHK18A1 and PHK18B1 similarly exhibited lower levels of E6 in passages 5 and 30 compared to those of passage 15 ( Figure 2B). Unfortunately, under our conditions, we were unable to detect the E6*I protein in any of the samples. p53 levels inversely correlated with those of E6 in HPV-18 transduced PHKs ( Figure 2C).
HR-HPVs E7 represses the tumor suppressor pRb leading to p16 overexpression in cervical tumors [31]. Here, HPV-18 E7 protein levels were indirectly measured by accessing the cellular targets pRb, p16, and p21. We observed that pRb levels were reduced in immortalized PHK18A1 and PHK18B1 (p30), in comparison with early passage cells (p5) ( Figure 2D). p16 and p21 levels were also comparable among variants ( Figure 2E,F).  those of E6 in HPV-18 transduced PHKs ( Figure 2C). HR-HPVs E7 represses the tumor suppressor pRb leading to p16 overexpression in cervical tumors [31]. Here, HPV-18 E7 protein levels were indirectly measured by accessing the cellular targets pRb, p16, and p21. We observed that pRb levels were reduced in immortalized PHK18A1 and PHK18B1 (p30), in comparison with early passage cells (p5) ( Figure 2D). p16 and p21 levels were also comparable among variants ( Figure 2E,F). Figure 2. E6, E6*I, E7, p53, p16, pRb, and p21 levels in PHK18A1 and PHK18B1: (A) qRT-PCR to quantify E6 + E6*I, E6 alone, and E7 expression. Mitochondrial ribosomal RNA S18 gene expression was used for normalization. Means and standard errors of three independent assays performed in triplicate is presented; (B) E6 protein levels. A total of 120 µg of protein extracts were fractionated in SDS-PAGE and the levels of E6 were analyzed by Western blotting; (C) p53, (D) pRb, (E) p16, and (F) p21 protein levels. A total of 80 µg of protein extracts were fractionated in SDS-PAGE and protein levels were analyzed by Western blotting. For (B-F), tubulin levels were used as loading control. Representative experiments of two assays are presented. Figure 2. E6, E6*I, E7, p53, p16, pRb, and p21 levels in PHK18A1 and PHK18B1: (A) qRT-PCR to quantify E6 + E6*I, E6 alone, and E7 expression. Mitochondrial ribosomal RNA S18 gene expression was used for normalization. Means and standard errors of three independent assays performed in triplicate is presented; (B) E6 protein levels. A total of 120 µg of protein extracts were fractionated in SDS-PAGE and the levels of E6 were analyzed by Western blotting; (C) p53, (D) pRb, (E) p16, and (F) p21 protein levels. A total of 80 µg of protein extracts were fractionated in SDS-PAGE and protein levels were analyzed by Western blotting. For (B-F), tubulin levels were used as loading control. Representative experiments of two assays are presented.

PHK18A1 and PHK18B1 Similarly Overcome Actinomycin-D (AD)-Induced Growth Arrest
We observed that AD treatment induced a significant decrease in the percentage of PHK18B1 in G0/G1 and an increase in the rate of PHK18A1 cells in the S phase ( Figure 3A). Interestingly, the sub-2N ratio between AD treated and untreated cells was significantly higher for PHK18B1, compared to PHK18A1 (p = 0.04), indicating that PHK18B1 were more prone to suffer from AD-induced cell death. Thus, we accessed the levels of caspases 3 and 7 in these cells. Using this approach, we observed no difference in the rate of apoptotic cells between untreated PHK18A1 and PHK18B1 (p = 0.35) ( Figure 3B). However, corroborating cell cycle analysis, following AD treatment the ratio of apoptotic cells was significantly higher for PHK18B1 (54.7%), compared to PHK18A1 (37.1%) (p = 0.00). Taken together, our data indicate that both PHK18A1 and PHK18B1 were capable of overcoming growth arrest induced by AD treatment, although PHK18B1 were significantly more prone to apoptosis following treatment. caspases 3 and 7 in these cells. Using this approach, we observed no difference in the rate of apoptotic cells between untreated PHK18A1 and PHK18B1 (p = 0.35) ( Figure 3B). However, corroborating cell cycle analysis, following AD treatment the ratio of apoptotic cells was significantly higher for PHK18B1 (54.7%), compared to PHK18A1 (37.1%) (p = 0.00). Taken together, our data indicate that both PHK18A1 and PHK18B1 were capable of overcoming growth arrest induced by AD treatment, although PHK18B1 were significantly more prone to apoptosis following treatment.

PHK18A1 and PHK18B1 Differ in Colony Formation Ability in Monolayer Cultures
The ability of immortalized PHKs in forming colonies in monolayer cell culture was accessed using two approaches: clonogenic assay and capacity to resist differentiation induced by serum and calcium. Although not statistically significant, PHK18A1 induced the growth of a larger number of colonies (average of 52 colonies) when compared to PHK18B1 (average of 36.7 colonies) (p = 0.13), ( Figure 4A). It is noteworthy that colonies formed by PHK18B1 were less cohesive and composed of a reduced number of cells, in

PHK18A1 and PHK18B1 Differ in Colony Formation Ability in Monolayer Cultures
The ability of immortalized PHKs in forming colonies in monolayer cell culture was accessed using two approaches: clonogenic assay and capacity to resist differentiation induced by serum and calcium. Although not statistically significant, PHK18A1 induced the growth of a larger number of colonies (average of 52 colonies) when compared to PHK18B1 (average of 36.7 colonies) (p = 0.13), ( Figure 4A). It is noteworthy that colonies formed by PHK18B1 were less cohesive and composed of a reduced number of cells, in comparison to those of PHK18A1, which originated more dense colonies of firmly grouped cells. Indeed, the average area of PHK18A1 (0.25 mm 2 ) colonies was significantly higher than that of PHK18B1 (0.16 mm 2 ) (p = 0.00).
Terminal differentiation was induced by maintaining cells in DMEM supplemented with 10% FBS and 1 µg/mL of hydrocortisone. We observed that under these conditions, PHK18B1 showed a greater ability to abrogate differentiation in this condition, compared to PHK18A1 (p = 0.00) ( Figure 4B).

PHK18A1 and PHK18B1 Differ in the Ability to form Colonies in 3D Cultures
Initially, cells were plated in low-attachment plates and spheroids formation and growth were monitored. After 14 days, we observed no differences in the number of spheroids formed by PHK18A1 (average of 12.6 spheroids), compared to PHK18B1 (average of 11.9 spheroids) (p = 0.70) ( Figure 4C). However, similar to what we observed in the monolayer ( Figure 4A), spheroids derived from PHK18B1 were also much less cohesive when compared to those of PHK18A1. The mean area of PHK18A1 derived spheroids (0.28 mm 2 ) was significantly higher than those resulting from PHK18B1 (0.09 mm 2 ) (p = 0.01).  We further seeded cells in low melting soft agar and after 15 days colonies were stained. Consistent with the clonogenic assay ( Figure 4A), PHK18A1 were more prone to form colonies compared to PHK18B1 (mean number of colonies PHK18A1 versus PHK18B1: 13.4 versus 8.7; p = 0.10) ( Figure 4D). Nevertheless, in this assay, the average colonies area was similar among variants (p = 0.67).
Finally, the ability of cells in forming spheroids was evaluated using a recently developed magnetic cell bioprinting technology (n3D Biosciences). Using this assay, a single spheroid per well is formed. Spheroids formed by PHK18B1 had a porous aspect, similar to parental PHKs but very different from those derived from PHK18A1 or HeLa ( Figure 4E). Even though these results originate from a single experiment in which an average of 20 replicates was performed for each cell line studied, they corroborate and expand data obtained in monolayer and the other 3D culture assays ( Figure 4A-D).

PHK18A1 and PHK18B1 Differ in Invasion but Not in Migration Ability
Wound healing assays were performed using mitomycin C treated parental cells. No significant differences were observed in cell cycle arrest induced by mitomycin C, neither between PHK18A1 and 18B1 (p = 0.65) nor between parental PHKs and PHK18A1 (p = 0.84) or parental PHKs and PHK18B1 (p = 0.50) (Supplementary Figure S1A). At 12 h following wounding, no differences in migration between PHK18A1 and PHK18B1 were observed (p = 0.28) ( Figure 5A). However, we detected important differences regarding cell migrations patterns among HPV-18 variants (videos in Supplementary Videos S1-S6). While PHK18A1 migrated collectively, i.e., we observed a group of cells migrating together toward the wound area, PHK18B1 cells tended to move more individually. Regarding invasion potential, PHK18A1 (mean absorbance 0.99) had a significantly higher invasion potential than PHK18B1 (mean absorbance 0.81) (p = 0.00). Moreover, while radial and central invasion was observed for PHK18A1, PHK18B1 exhibited solely radial invasion ( Figure 5B). The invasion of large clusters of PHK18A1 throughout inserts was also observed, whereas PHK18B1 invaded in small groups of cells.

PHK18A1 and PHK18B1 Spheroids Differ in Epithelial-Mesenchymal Transition (EMT) Phenotype
Epithelial to mesenchymal transition (EMT) involves ECM degradation and is characterized by loss of E-cadherin concomitant with augmented vimentin levels [32]. To evaluate whether the greater invasion ability of PHK18A1 was associated with EMT, the levels of both proteins were assessed among spheroids formed after plating in low-attachment plates. We observed that parental cells presented an epithelial phenotype (more cells expressing E-cadherin), while immortalized PHKs harbored a mesenchymal phenotype as inferred by the higher number of cells expressing vimentin. Significant higher levels of cells expressing vimentin were observed among PHK18A1 (42.8%), in comparison to PHK18B1 (22.3%) (p < 0.00) ( Figure 5C, Supplementary Figure S2).

PHK18A1 and PHK18B1 Form Similar Epithelia in Raft Cultures
PHK18A1 and PHK18B1 were used to establish organotypic raft cultures that allow for differentiation similar to that observed in epithelial tissues [33]. Ordered stratification reminiscent of the normal skin was observed in raft cultures derived from parental PHKs, where we observed basophilic undifferentiated basal cells, in addition to several layers of eosinophilic suprabasal cells ( Figure 6). On the other hand, tissues obtained from PHK18A1 and PHK18B1 were hyperplastic and presented abnormal stratification preventing the discrimination between basal and suprabasal layers. No differences were observed among cultures derived from PHK18A1 and PHK18B1 concerning epithelial thickness or morphology. uate whether the greater invasion ability of PHK18A1 was associated with EMT, the levels of both proteins were assessed among spheroids formed after plating in low-attachment plates. We observed that parental cells presented an epithelial phenotype (more cells expressing E-cadherin), while immortalized PHKs harbored a mesenchymal phenotype as inferred by the higher number of cells expressing vimentin. Significant higher levels of cells expressing vimentin were observed among PHK18A1 (42.8%), in comparison to PHK18B1 (22.3%) (p < 0.00) ( Figure 5C, Supplementary Figure S2).  CK10 levels were uniformly detected throughout cells of all suprabasal epithelium layers of tissues derived from parental PHKs or from immortalized PHKs, regardless of the HPV-18 variant. CK10 is a marker of differentiation, as well as of hyperplasia reserve cells, squamous metaplasia, and the cervical transformation zone [34]. No differences were observed in p16 or PCNA staining patterns between PHK18A1 and PHK18B1 ( Figure 6). It is noteworthy that in tissues derived from PHK18A1 and PHK18B1 high levels of CK10 and PCNA were concomitantly detected in most cells within suprabasal layers, indicating that E6 and E7 proteins of the two HPV-18 variants were capable of inducing both proliferation and differentiation in the same cell. It is noteworthy, however, that PCNA levels seem to be higher in PHK18B1 rafts, in comparison to PHK18A1 derived tissue. Taken together, our data show no significant differences regarding morphology and levels of the different proteins analyzed between PHK18A1 and PHK18B1-derived epithelia that could indicate lesions with different tumorigenic potential.
the HPV-18 variant. CK10 is a marker of differentiation, as well as of hyperplasia reserve cells, squamous metaplasia, and the cervical transformation zone [34]. No differences were observed in p16 or PCNA staining patterns between PHK18A1 and PHK18B1 (Figure 6). It is noteworthy that in tissues derived from PHK18A1 and PHK18B1 high levels of CK10 and PCNA were concomitantly detected in most cells within suprabasal layers, indicating that E6 and E7 proteins of the two HPV-18 variants were capable of inducing both proliferation and differentiation in the same cell. It is noteworthy, however, that PCNA levels seem to be higher in PHK18B1 rafts, in comparison to PHK18A1 derived tissue. Taken together, our data show no significant differences regarding morphology and levels of the different proteins analyzed between PHK18A1 and PHK18B1-derived epithelia that could indicate lesions with different tumorigenic potential.

Discussion
We investigated E6/E7 functional characteristics relevant to carcinogenesis of two HPV-18 sublineage variants commonly detected in our population [6] in the genetic background of natural host cells using an in vitro model that resembles persistent infection. PHK18A1 reached p30 at a significantly earlier time point, compared to PHK18B1, and an evident crisis prior to immortalization was not observed. Our data are in line with those of Lace et al. [35], who also did not observe a clear crisis pre-immortalization period when PHKs of different anatomical origins were infected with HPV-16 or HPV-18 whole genome but contrasts to previous results from our group and others regarding PHKs immortalized by E6 and E7 of different HPV-16 variants in which a clear "crisis" was observed between p10-15 [7,8].
During HPV-induced carcinogenesis, HR-HPV E7 degrades pRb, releasing the E2F transcription factor that induces cell proliferation [18]. Furthermore, HR-HPV E6 induces the degradation of p53 and activates hTERT [36], which are crucial for infected cells to abrogate growth arrest and apoptosis induced by DNA damage, and to extend the lifespan of PHKs, respectively. E6 protein levels fluctuated during passages in culture in both PHK18A1 and PHK18B1, and p53 levels were inversely correlated to those of E6. As HPV-18 E6*I has been shown to inhibit E6-mediated degradation of p53, we hypothesize that in our model, although E6*I mRNA levels did not vary between variants, at p30 in both PHK18A1 and PHK18B1, E6*I protein levels are higher than full-length E6, which, however, can direct degradation of other cellular proteins in the absence of full-length E6 protein to sustain immortalization [37]. Actually, although p53 and pRB are the most studied interactors of E6 and E7, respectively, many other cellular targets of both viral proteins reported [38] can respond to the differences observed between PHK18A1 and PHK18B1 throughout this study. Unfortunately, under our experimental conditions, we were unable to detect E6*I protein in any of the samples, limiting conclusions. Our data contrast with a previous report in which high levels of E6*I transcripts were detected solely in MCF-7 and C33 cells transfected with E6 of an HPV-18 B1 sublineage variant [23,24]. Discrepancies observed between studies most probably rely upon differences regarding the models and the genetic background of the cells used, hindering any further consideration.
The induction of growth arrest is an important cellular response to DNA damage which avoids accumulation of mutation driving carcinogenesis [39]. Both HPV-18 variants were able to overcome AD-induced growth arrest, once the rate of cell in the proliferative S phase increased after treatment. AD has also been shown to be a potent inducer of apoptosis [40]. The higher levels of active caspases 3 and 7 and sub-2N cells observed in AD treated PHK18B1 indicate that PHK18A1 were more resistant to AD-induced apoptosis. This property may confer to HPV-18 A1 sublineage variant a higher oncogenic potential once reduced susceptibility to apoptosis supports tumor growth [41].
The ability of single cells in forming colonies through clonal expansion is a largely used measure of oncogenic potential [42]. Some approaches were used here and overall show that PHK18A1 were able to form a higher number and larger colonies and spheroids in monolayer and 3D cultures, respectively. Of importance, our study highlights qualitative differences among PHKs immortalized by the different variants: in contrast to PHK18A1, which formed more compact colonies and spheroids of firmly grouped cells, PHK18B1 tended to form structures of noncohesive cellular clusters, which resembles colonies and spheroids formed by parental PHKs. We hypothesize that interactions of E6/E7 of the different HPV-18 variants with different cellular proteins, for instance, PTEN [25], may explain the differences observed. Although in a different genetic background, it was also previously observed a higher number of colonies arising from NIH3T3 cells expressing solely E6 of the HPV-18 A1 variant grown in soft agar, compared to other variants [23]. Although migration was similar among variants, migration patterns were also different: PHK18A1 tended to migrate in clusters of cells as commonly observed in epithelial tumors, whereas PHK18B1 migrated individually similar to solid stromal tumors. In fact, a broad spectrum of migration and invasion mechanisms have been observed among cancer cells [32]. It was hypothesized that the advantage of migrating in clusters over migrating individually confers protection of inner cells from immunological attack, thereby increasing the efficiency of tumor invasion and survival. Nevertheless, further studies are needed to identify the mechanisms underlying each type of cell migration observed here. Our data contrast to that of Fragoso-Ontiveros et al. [43], who observed significantly higher migration potential of cells expressing only E6 the HPV-18 B1 variant, compared to the A1 counterpart. However, direct comparisons are also hindered by differences in cell assay models.
PHK18A1 showed an enhanced capacity for invasion through a collagen matrix when compared to PHK18B1. We presume that the interaction of viral and target proteins that compose the basement membrane, interstitial stroma, and/or extracellular matrix may contribute to invasion during HPV-induced carcinogenesis. For instance, it was shown that the HPV-16 E6 PDZ binding motif significantly contributes to Caski and HeLa cells migration and invasion [44]. We suggest that the higher and lower number of cells expressing vimentin and E-cadherin, respectively, observed in PHK18A1 may contribute, at least in part, to the higher invasion potential observed for these cells, once downregulated E-cadherin is associated with decreased cell-cell adhesion, facilitating EMT and invasion into adjacent tissues [32].
E6 expression is also crucial for PHKs to resist terminal differentiation induced by serum and calcium [45]. We observed that, when cells were maintained in a differentiationinducing medium, PHK18B1 exhibited greater resistance to differentiation, compared to PHK18A1. However, we may not exclude the chance that these cells, by a mechanism not elucidated here, have a higher ability to survive when grown in a calcium-rich medium. Furthermore, once MTT, which was used to visualize colonies, is processed by the mitochondria, the possibility exists that E6/E7 of the HPV-18 variants analyzed differently impact cell respiration. Although some studies have evaluated the role of E6 and E6*I in mitochondrial activity associated with oxidative stress and cell death, there are no reports regarding the influence of mitochondrial functions on the differentiation of HPVinfected keratinocytes [23,46]. It should be noted also that resistance to differentiation and immortalization may develop by independent pathways.
Finally, we did not observe any evident differences between PHK18A1 and PHK18B1 in forming epithelia in organotypic culture: both showed hyperplasia, in addition to high levels of CK10, PCNA, and p16 throughout the tissue, compatible with cervical intraepithelial neoplasia [20,47]. Nevertheless, it is feasible that the different parameters evaluated were not sufficient to reveal lesions with different degrees of aggressiveness. Furthermore, although the expressions of HR-HPV E6 and E7 induce alterations in keratinocyte differentiation in organotypic cultures [47][48][49], this model may not be adequate to unravel differences attributable to these variants regarding association to different histological subtypes of CC. In fact, although PHKs have been extensively used in the last decades in functional studies aiming to understand the mechanisms of HPV-induced cervical carcinogenesis better, we believe that pathological differences between cervical SCC and ADC would be better observed in the context of squamocolumnar junction cells, which are believed to constitute the source of cervical cancer [50].
HPV-18 naturally occurring variants might exhibit other biological differences than those evaluated in this study, which may affect their pathogenic potential, including differences in the host immune response. However, as a whole, our data indicate that PHK18A1 exhibits a phenotype that more closely resembles transformed cells, in comparison to PHK18B1. Further studies are necessary to understand the underlying biochemical mechanisms and pathways behind the observed differences.

Conclusions
PHKs immortalized by E6/E7 of HPV-18 A1 and B1 variants were characterized. Regarding hallmarks of cancer cells, PHK18A1 showed higher oncogenic potential. Furthermore, we highlight qualitative differences among immortalized PHKs, which might impact their association to different CC histological subtypes.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.