Pten loss in Lgr5+ hair follicle stem cells promotes SCC development

Accumulating data support that tissue stem cells give rise to cancer cells. Hair follicle stem cells (HFSCs) undergo cyclic quiescence and activation and may sever as the origin of cutaneous squamous cell carcinoma (SCC). Pten is a tumor suppressor gene that is frequently mutated in hereditary cancer syndromes such as Cowden disease, which is featured with papillomatosis in cutaneous tissues and hyperkeratosis in the acral region of the skin. Additionally, mice with keratinocyte-specific Pten deficiency (k5-Pten-/- mice) show epidermal hyperplasia and spontaneous tumor formation. However, the impact of Pten mutation in HFSCs, such as in Lgr5+ HFSCs, on SCC formation is unclear. Methods: We established experiments with wildtype and Lgr5-CreER; Ptenflox/flox mice, and used DMBA/TPA two-stage skin carcinogenesis model to explore the effect of Pten loss in Lgr5+ HFSCs of 3 weeks old mice in skin carcinogenesis. In vitro experiments (cell culture and protein expression analysis) are employed to investigate molecular mechanisms involved. Results: Pten loss in Lgr5+ HFSCs promoted SCC formation, which was attenuated in TNF-/- mice. Notably, β-catenin loss in Lgr5+ HFSCs decreased the formation of SCC. In addition, Pten loss in cultured epidermal stem cells upregulated the levels of both phospho-Akt and β-catenin. Conclusion: Pten loss in Lgr5+ cells induced Akt/β-catenin signaling, and SCCs can subsequently be raised as progeny from these primed Lgr5+ stem cells.


Introduction
Accumulating evidence indicates that certain cancers such as leukemia arise from somatic stem cells with gene mutations [1,2]. Lgr5 marks hair follicle stem cells (HFSCs) located in the lower bulge and the secondary hair germ of the telogen hair follicle (HF) [3]. Lgr5 + cells substantially contribute to the cycling portion of anagen HFs and take part in the reepithelialization in skin wound healing [3,4]. In addition, HFSCs with misactivated Hedgehog (Hh) signaling recruited to the epidermis by wounding contribute to basal cell carcinoma (BCC)-like lesions [5,6]. Moreover, progeny of Lgr5 + HFSCs contribute to papillomavirus-induced SCC, the second most common skin cancer [7]. Pten (phosphatase and tensin Ivyspring International Publisher homolog deleted on chromosome ten), is a tumor suppressor gene that frequently mutated in hereditary cancer syndromes such as Cowden disease, which is featured with papillomatosis in cutaneous tissues and hyperkeratosis in the acral region of the skin [8,9], and many other cancers [10]. Additionally, mice with keratinocyte-specific Pten deficiency show epidermal hyperplasia and spontaneous tumor formation [11]. However, the impact of Pten mutation in HFSCs, such as in Lgr5 + HFSCs, on cutaneous squamous cell carcinoma (SCC) formation is unclear. β-catenin has also shown to be involved in the development of SCC. SCC exhibit a preferential nuclear location of β-catenin, and inhibition of β-catenin signaling significantly attenuates the growth of SCC cells [12][13][14]. However, whether the activity of β-catenin signaling in HFSCs affecting SCC formation is unclear. In addition, it is also desired to uncover the interaction between Pten/Akt and β-catenin signaling in SCC formation. Furthermore, previous studies show that TNF -/mice are resistant to DMBA/TPA-induced skin SCC [15], suggestive of the involvement of TNF in SCC development. In this study, we found that mice with Pten knock out in Lgr5 + HFSCs showed increased incidences of skin papilloma and SCC upon DMBA/TPA induction, while double loss of Pten /CTNNB1(β-catenin gene) or Pten/TNF in Lgr5 + HFSCs greatly diminished the tumorigenesis. Thus our data indicate that Pten loss in HFSCs greatly promotes the formation of SCCs, and β-catenin and TNF are critically involved.

Mice
C57BL/B6 mice (6-week-old, female) were purchased from Guangdong Medical Laboratory Animal Center, Guangzhou, China. Lgr5-GFP-Cre-ERT2 (Lgr5-CreER) mice were obtained from Jackson Laboratory (Stock No.: 008875). The mice were crossed with Pten flox/flox mice (a gift from Dr. Hong Wu at the University of California, Los Angeles) to obtain Lgr5-CreER;Pten flox/flox mice, whose identify were verified ( Figure S1A-B To knock out Pten, β-catenin, in Lgr5 cells, Lgr5-CreER; Pten flox/flox mice, Lgr5-CreER; Pten flox/flox ;β-catenin flox/flox mice, Lgr5-CreER; Pten flox/flox ; TNF KO (knock out) mice, Lgr5-CreER; Pten flox/flox ; Rosa-mTmG mice aged 3 weeks received an intraperitoneal injection of 100 μL of tamoxifen (TAM, Sigma Aldrich) in corn oil at a concentration of 10 mg/mL for three times. In addition, we used littermate mice for control in all genetic mice model involved experiments. Mice were randomly divided into groups using a random-number table. The animals were maintained in a temperature-controlled environment (20 ± 1 °C) with free access to food and water. All procedures were performed with the approval of Animal Ethics Committee of Shenzhen Center for Disease Control and Prevention (CDC).

Tumor induction in mice
Skin SCC in mice was induced as previously described [16,17]. Briefly, 25 μg DMBA (Sigma Aldrich) in 200 µL acetone was applied to the dorsal skin after shaving. After 2 weeks, TPA (10 nmol) in 200 μL was applied to the same area twice weekly for up to 30 weeks. Skin specimens were collected 4 weeks and 9 weeks after DMBA treatment, and when papilloma and SCC formed. The number of tumors per mouse was counted each week as palpable mass >1 mm in size. Tumor volume was also estimated and recorded periodically [16,18,19].

Immunofluorescence (IF) staining
Freshly obtained skin samples from mice back with hair removal were fixed in 4% paraformaldehyde for 8 h. Then were taken off water in 10%, 20% and 30% sucrose gradient for 8 h and embedded in Tissue Freezing Medium (SAKURA Tissue-Tek® OCT Compound). Frozen tissue sections of the skin were incubated with different primary antibodies at 4 ℃ overnight, which were anti-p-Akt Nuclei were stained with 4, 6-diamidino-2-phenylin-dole (DAPI). After mounting, samples were visualized under confocal microscope (FV1000; Olympus, Tokyo, Japan) [20].

Isolation and culture of epidermal stem cells
Neonatal mouse dorsal skin was harvested from Pten flox/flox ;Rose-mTmG mice 1~3 days after birth. The tissue was cut into 2~3 mm 2 pieces, washed 3 times in HBSS, and digested with 0.3% Dispase II (sigma) for 90 min at 37 °C. The epidermis was manually removed from the tissue. Epithelial stem cells were isolated based on their high adhesive property [21]. Briefly, the epithelial layer was cut into slurry and treated with 0.2% collagenase I (Sigma) for 60 min at 37 °C with shaking and filtered through a 40 µm cell strainer. The cells were seeded in tissue culture dishes coated with 50 µg/mL collagen I (Sigma) and incubated in CnT-07 PCT Epidermal Keratinocyte Medium (CnT-07; CELLnTEC Advanced Cell Systems, Bern, Switzerland) with supplements provided by the manufacture for 60 min. The non-adherent cells were removed and the adherent cells were maintained. When reaching 80% confluence, the culture was passaged after digestion with accutase (Sigma) [4].

Transfection
Epidermal stem cells were seeded on 6 wells culture dishes, and when cells reached 40% confluence, the medium was exchanged with 1 mL new medium containing 40 µL 1.0×10 10 pfu GFP-Cre-adenovirus and incubated for 24 h. Then the medium was changed to growth medium, and cells were collected after incubation for 24, 48 and 72 h.

Histological analysis
Freshly obtained skin samples from mice back with hair removal were fixed in 10% formalin or other fixatives for 12-24 h at room temperature. After dehydration, tissues were embedded in paraffin. Tissue sections were rehydrated with 100% ethanol, 95% ethanol, 75% ethanol, and deionized H2O, 3 min each. Then use hematoxylin to stain nuclear and eosin to stain cytoplasm and visualized with a Leica microscope. The mean width of HFs was measured, and 100 HFs per mice were measured per mouse.

Statistical analysis
Results are expressed as mean±s.e.m. unless stated otherwise. Statistical comparisons between two groups were evaluated by Student's t-test. A probability (P) value <0.05 was considered to indicate statistical significance.

Pten loss in Lgr5 + stem cells induces hair follicle hyperplasia
A previous study showed that Pten deletion in Lgr5 + HFSCs in 7∼8 weeks old mice did not induce the hyperproliferation of HFSCs [22]. Here we examined the influence of Pten loss on Lgr5 + HFSCs in 3 weeks old Lgr5-CreER;Pten flox/flox mice. The mice were treated with tamoxifen to induce Pten loss in Lgr5 + cells (Lgr5-Pten -/-). At the 12 th day after intraperitoneal injection of tamoxifen, immunofluorescence staining of the skin tissue showed that Pten level was extensively reduced in Lgr5 + HFSCs, resulting in increased levels of p-Akt ( Figure S2A-B). Notably, 20 days after tamoxifen administration, the mice showed significantly enlarged HF containing more cells (hyperplasia), compared to Lgr5-CreER;Pten flox/flox mice without tamoxifen treatment (Lgr5-Pten +/+ ) ( Figure  1A, C). The enlarged HF shortened when proceeding into the telogen phase but still much wider than in the Lgr5-Pten +/+ mice, as examined 40 days (8-9 weeks of age) after Pten deletion ( Figure 1B, D). Taken together, the data indicate that Pten loss in Lgr5 + HFSCs induce HF hyperplasia in 3 weeks old mice.

Pten loss in Lgr5 + stem cells promotes papilloma formation
To gain insight into the role of Pten loss in Lgr5 + HFSCs in the development of SCCs, we employed DMBA/TPA SCC mouse model [16,17]. Six weeks after Pten deletion in Lgr5-Pten -/mice, DMBA was applied to the dorsal skin of mice after hair shaving, followed by TPA treatment (Figure 2A). Lgr5-Pten +/+ received equal treatment of DMBA/TPA were used as control. Four weeks after the DMBA treatment, the dorsal skin of mice showed hyperplasia in both Lgr5-Pten -/mice and Lgr5-Pten +/+ mice, whereas the epidermal layer of the skin was much thicker in Lgr5-Pten -/mice ( Figure 2B). Nine weeks after DMBA treatment, significant hyperplasia was observed in both groups, but the lesion was more severe in Lgr5-Pten -/mice ( Figure 2B). Importantly, skin papillomata were observed in Lgr5-Pten -/mice early in the 5 th week after DMBA treatment, whereas in Lgr5-Pten +/+ mice papillomata were found 8 weeks after DMBA treatment. In addition, the incidence of mice developing papillomata was higher in Lgr5-Pten -/mice (100%, n=8) than in Lgr5-Pten +/+ mice (85.7%, n=7) 25 weeks after TPA treatment ( Figure 2C, D). Together, these data indicate that Pten knockout in Lgr5 + HFSCs greatly promotes skin papilloma formation.

Pten loss in Lgr5 + HFSCs promotes SCC development
We next examined whether Pten knockout in Lgr5 + HFSCs promoted skin SCC development. We found that some papillomata in Lgr5-Pten -/mice developed into invasive lesions 18 weeks after DMBA treatment, where proliferating cells in the epidermisand follicle-like structures invaded into the space between them in histology, exhibiting the feature of SCC; by contrast, similar lesions were found in much later in Lgr5-Pten +/+ mice (26 weeks after DMBA treatment) ( Figure  3A-B). 26 weeks after DMBA treatment, 62.5% of Lgr5-Pten -/mice developed SCC lesions compared to 14.3% in Lgr5-Pten +/+ mice (Figure 3  A). These data indicate that Pten deletion in Lgr5 + HFSCs contributes to the development of SCC.

Lineage tracing of Lgr5 + HFSCs and their progeny in tumor development
To further examine the contribution of Lgr5 + HFSCs with Pten loss and their progeny to skin tumor development, we used Lgr5-CreER; Pten flox/flox ; Rosa-mTmG (Lgr5-Pten -/--mTmG) and Lgr5-CreER;Rosa-mTmG (Lgr5-mTmG) mice (control). The mice were treated with DMBA/TPA as indicated in Figure 2A. The result showed that an increased amount of cells were progenies of Lgr5 + stem cells (membrane tdTomato -/ membrane GFP + , mT -/mG + ) in hyperplastic HFs and epidermis in Lgr5-Pten -/--mTmG mice, whereas in Lgr5-mTmG mice the HF and epidermis did not show hyperplasia and contained fewer mT -/mG + cells ( Figure 4A, B). With progression of the disease, we found mT -/mG + cells in papillomata and in SCC, which formed colonies, in Lgr5-Pten -/--mTmG mice, but not in Lgr5-mTmG mice ( Figure 4C, D). To verify the contribution of Lgr5 + cells (with Pten loss) in papilloma and SCC, we performed IF staining of tissue sections of these tumors derived from Lgr5-Pten -/mice and Lgr5-Pten +/+ mice for the expression of Lgr5, and we did not detect the presence of Lgr5-expressing cells ( Figure S3). These data suggest that the progeny of Lgr5 + HFSCs with Pten loss contribute to skin papilloma and SCC.

β-catenin effects at the downstream of Pten/Akt signaling
To elucidate the molecular mechanisms underlying the role of Pten loss in Lgr5 + HFSCs in SCC development, we performed in vitro and in vivo experiments. Immunostaining of the skin tissue of Lgr5-Pten -/mice showed elevated expression levels of p-β-catenin (Ser552) in the HF and epithelial cells ( Figure 5A), indicating increased β-catenin signaling after Pten loss. To gain more insight into the crosstalk between Pten/Akt and β-catenin in epidermal stem cells (Epi-SCs), Epi-SCs were isolated from Pten flox/flox ;Rose-mTmG mice and the cells were treated with Ad-Cre virus ( Figure S4) to induce the knockout of Pten. Western blot showed that the expression of Pten decreased markedly at the 48 h and 72 h after Ad-Cre virus treatment, meanwhile the level of p-Akt increased, but the level of total Akt unchanged. In addition, increased levels of p-Gsk-3β (Ser9) and p-β-catenin (Ser552), but unchanged levels of p-β-catenin (Ser675) were also detected ( Figure 5B, C). Consistently, elevated level of p-Akt and p-β-catenin (Ser552) were detected in papillomata ( Figure 5D, E). Taken together, these data suggest Pten loss activates β-catenin possibly through the Akt mediated phosphorylation of β-catenin (Ser552). We detected more Ki67-expressing cells in papillomata of mice with Pten loss in Lgr5 + cells ( Figure S5). To verify the role of β-catenin in Pten loss induced SCC formation, we generated Lgr5-CreER;Pten flox/flox ;β-catenin flox/flox mice, whose Pten and β-catenin were deficient in Lgr5 + cells upon induction with tamoxifen (Lgr5-Pten -/--βcatenin -/mice); we found that less severe HF and epidermal hyperplasia in the mice compared to Lgr5-Pten -/mice ( Figure 6A, C). Moreover, the incidence of tumor formation and the average number of papillomata per mouse also decreased markedly in Lgr5-Pten -/--β-catenin -/mice ( Figure 6B, D, E). These data suggest that the loss of Pten in Lgr5 + HFSCs is likely through β-catenin signaling to promote SCC development.
Moreover, the average number of papillomata per mouse in Lgr5-Pten -/--TNF -/mice was much lower compared to that in Lgr5-Pten -/mice ( Figure 2D, Figure 7F), and TNF -/mice did not develop papilloma when receiving the same DMBA /TPA treatment ( Figure 7E-F). These data indicate that TNF has a crucial role in Lgr5 + HFSCs with Pten loss induced papilloma formation.

Discussion
Lgr5-expressing cells have been found in many tumors, such as colon cancer [26,27], papillary thyroid cancer [28], breast cancer [29], and gastric cancer [30]. Recently, Lgr5 has been found to induce epithelial-mesenchymal transition (EMT) in human hepatocellular carcinoma cells and inhibit their apoptosis, resulting in drug resistance [31]. In human skin SCC, high levels of Lgr5 have been found in immunohistological analysis [32]. In the skin of mice, Lgr5 + stem cells in the HF undergo active cyclic proliferation, and contribute to epidermal cells after wounding to the skin [3,4]; but lineage tracing of Lgr5 + cells did not show their direct contribution to skin SCC, nor their progeny [33,34]. White and colleagues found that gain of oncogene Ras or the loss of tumor suppressors p53 or Pten in HFSCs are unable to initiate tumors during the telogen phase of the HF in adult mice, suggesting that the mechanisms that keep HFSCs quiescent are dominant over the oncogenic influences. However, gain of Kras G12D , accompanied with loss of both Pten and p53 was sufficient to induce skin malignancies in quiescent HFSCs. These data support that Pten has a crucial role in maintaining quiescence in the presence of tumorigenic stimuli [22]. While it would be interesting to explore the role of Pten loss in skin malignancy development without other tumorigenic stimuli, Pten-only deletion in Keratin 15 + cells during HFSC quiescence did not result in hyperplasia in adult mice [28]. However, in this study, we found that Pten deletion in Lgr5 + HFSCs at the telogen/anagen transition of 3 weeks old mice induced hyperplasia in HFs, and the mice showed more severe hyperplasia in the epidermis, larger number of papillomata and increased incidence of SCC. These data support that Pten deletion alone in the activated HFSCs is sufficient to drive these stem cell into hyperplasia. Our data further support that Pten loss induced SCC development is dependent on the activation of HFSCs. Lgr5 + HFSCs are the first activated stem cell population in response to anagen-initiating signals, whereas keratin 15 + HFSCs remain quiescent even in anagen phase [3]. This may explain, at least in part, why Pten deletion alone in Lgr5 + HFSCs is sufficient to drive HFs to hyperplasia, whereas a combination of Kras G12D and the loss of both Pten and p53 are needed to derive the Keratin 15 + HFSCs into hyperplasia. However, more concrete studies should be performed to deeply uncover the differential response of stem cells in activation and in quiescence to tumorigenic stimuli.
Wnt/β-catenin exhibits complicated effect on both the development of stem cells and cancers [35][36][37][38]. In the HF, the transit-amplifying matrix compartment appears to be the target for malignant transformation by mutational activation of the Wnt cascade [39]. Constitutive activation of β-catenin in HFs induces pilomatricoma-like lesions, where the exterior zone of the tumor is composed of densely packed cells resembling the matrix of HFs [40]. Another HF tumor, namely trichofolliculoma, was developed in tamoxifen-inducible β-catenin transgene mice [41]. Notably, the most spontaneous pilomatricomas in humans are featured with activating mutations in β-catenin [42]. Recent evidence has shown that in many organs, such as the HF, resident adult stem cells can be the initiating cells for cancers [43,44]. In this study, deletion of both Pten and β-catenin in Lgr5 + HFSCs result in reduced HF hyperplasia and SCC development. Together with the in vitro data, our study support that the effect of β-catenin is downstream of Pten/Akt signaling in the initiation and development of SCC. However, more studies with gain and lose function of β-catenin and Pten/Akt signaling in HFSCs are desired, to further explore the crosstalk between Pten/Akt and β-catenin and its role in skin cancer development. In addition, our data suggest that TNF plays an essential role in Pten loss in HFSCs induced hyperplasia and SCC development. However, the relationship between TNF and Pten/Akt signaling in HFSCs activation and SCC development was still largely unknown. Given the evidence of TNF as a key component of HFSCs activation [4,45], TNF and Pten/Akt signaling pathways possibly exhibit synergistic effect on SCC development via activating and transforming HFSC, respectively.
Conclusively, our study provide evidence that Pten loss in activating Lgr5 + HFSCs was sufficient to drive HFs into hyperplasia, and the Pten loss in Lgr5 + HFSCs plays an important role in subsequent SCC formation. Moreover, we found both β-catenin and TNF play essential roles in HFSCs with Pten loss induced hyperplasia and SCC development. Together, these data indicate that Pten loss in HFSCs is a key driver in SCC initiation and development, where active interaction of Pten/Akt signaling with other signals such as β-catenin and TNF are required.