Lkb1 Loss Promotes Tumor Progression of BRAFV600E-Induced Lung Adenomas

Aberrant activation of MAP kinase signaling pathway and loss of tumor suppressor LKB1 have been implicated in lung cancer development and progression. Although oncogenic KRAS mutations are frequent, BRAF mutations (BRAFV600E) are found in 3% of human non-small cell lung cancers. Contrary to KRAS mutant tumors, BRAFV600E-induced tumors are benign adenomas that fail to progess. Interestingly, loss of tumor supressor LKB1 coexists with KRAS oncogenic mutations and synergizes in tumor formation and progression, however, its cooperation with BRAFV600E oncogene is unknown. Our results describe a lung cell population in neonates mice where expression of BRAFV600E leads to lung adenoma development. Importantly, expression of BRAFV600E concomitant with the loss of only a single-copy of Lkb1, overcomes senencence–like features of BRAFV600E-mutant adenomas leading malignization to carcinomas. These results posit LKB1 haploinsufficiency as a risk factor for tumor progression of BRAFV600E mutated lung adenomas in human cancer patients.


Introduction
Lung cancer is the most prevalent cancer in the industrialized world. Non-small-cell lung cancer (NSCLC) subtypes account for approximately 80-85% of all cases of lung cancer, where adenocarcinomas represent approximately 50% of all NSCLC. Despite its prevalence and characteristically high mortality rates the cellular and genetic origins of the disease are not completely understood.
Activation of RAS-RAF-MEK-ERK signaling pathway is implicated in the development of a wide variety of human tumors. Mutational activation of KRAS or BRAF has been detected in ,25% of NSCLCs. While the majority of these, harbor KRAS mutations, activating mutations in BRAF account for 3% of all NSCLCs [1,2,3]. In addition, somatic mutations in genes responsible for RAS pathway activation (i.e. EGF receptor or ERBB2) are detected in ,13% of human lung adenocarcinomas [4].
Several genetically engineered mouse models (GEM) of NSCLC have been generated. Two independent models demonstrate that expression of a mutant form of Kras (Kras G12D or Kras G12V ) from its endogenous promoter leads to the development of lung tumors that progress to adenocarcinomas [5,6]. Another model carrying a genetically modified allele of Braf (Braf CA ), which expresses wild type Braf prior to Cre-mediated recombination after which constitutively active BRAF V600E is expressed, elicited the growth of benign neoplastic adenomas [7]. A striking feature of the BRAF-mutant lung tumors of these animals is that they fail to progress to carcinoma and, instead, exhibit growth arrest and senescence-like features. This senescence-like phenotype could be overcome through concomitant mutation of p53 or p16 Ink4a / p19 Arf , which allow the tumors to progress to adenocarcinomas [7].
The mechanism behind the phenotypic differences between lung tumors expressing mutational activated KRAS or BRAF is not well understood. The early lesions induced by KRAS G12D and BRAF V600E suggest a common cell of origin expressing markers for alveolar type 2 (AT2) cells [8]. Nevertheless, phenotypic differences in later stages of tumor development leave open the possibility that these unique tumor types may have different cells of origin. In fact, several cell populations exhibiting stem-cell properties have been described in distinct anatomical regions of the lung [9,10,11,12].
Tumor suppressor STK11 (LKB1) gene is frequently deleted in lung tumors [13]. Nearly half of the NSCLCs harbor somatic and homozygous inactivating mutations in LKB1 [14], where it loss has been implicated in tumorigenesis and metastasis [5,15]. LKB1 loss has been observed concurrently with KRAS oncogenic mutations in human lung cancer, and animal models have confirmed the synergy of the combined mutations [5]. However the remaining question is whether LKB1 loss plays a role in BRAF V600E -driven tumors.
Here we show that in the course of modeling malignant melanoma in a inducible tyrosinase promoter transgenic system, activation of BRAF V600E oncogene in 2.5 days old mice by 4-OHtamoxifen administration, make them susceptible to develop lung adenomas. Interestingly, the senescence-like phenotype of BRAF V600E -induced adenomas could be overcome through concomitant deletion of LKB1 kinase gene allowing them to progress to carcinomas.

Mouse Strains and 4OHTx Treatment
Braf CA/CA strain has been previously described [7]

EYFP Visualization
Tyr::Cre ERT2 ; Braf CA/+ ; Lkb1 flox/+ mice were sacrificed at two, three and five days after treatment with 4OHTx. Tissue samples were fixed in 4% paraformaldehyde for 12 h and then cryoprotected in 30% sucrose in PBS o/n, followed by another incubation in 30% sucrose and 50% of OCT for additional 24 h. Then samples were frozen in OCT. Cryosections from these samples were used and EYFP native fluorescence was detected by microscopy.

Neonatal Activation of BRAF V600E Drives Aberrant Proliferation of Lung Cells
We modeled a genetic engineered mouse to study BRAF V600E and LKB1 cooperation in malignant melanoma. To this end, we used a conditional tyrosinase promoter transgenic system (Tyr::-Cre ERT2 ) and Braf CA mice [7], where Braf CA mice express wild type BRAF prior to Cre-mediated recombination, at which time oncogenic Braf V600E is expressed in physiological amounts [7]. After generation of Tyr::Cre ERT2 ; Braf CA/+ and Tyr::Cre ERT2 ; Braf CA/CA mice, recombination at the Braf locus was achieved by neonatal treatment with 4OH-tamoxifen at postnatal day 2.5 ( Figure 1A). During characterization of the mice's phenotype, the occurrence of lung adenomas in tyrosinase promoter-driven transgenic mice (Tyr::Cre ERT2 ) expressing BRAF V600E , prompted a detailed analysis of transgene expression in the lung. Examination of Cre-recombinase expression was performed in two and a half old day wild type (WT) (n = 3) and Tyr::Cre ERT2 ; Braf CA/+ mice (n = 3) treated with 4OHTx. Transgenic mice showed a nuclear expression of Cre-recombinase ( Figure 1B). The same mice were examined three days after 4OHTx treatment, and the activation of the BRAF downstream target ERK1/2 after expression of Braf CA was assessed detecting p-ERK1/2 amounts ( Figure 1C).
To further confirm the expression of Cre-recombinase in tyrosinase promoter-driven transgenic mice, we generated Tyr::-Cre ERT2 ; ROSA26-lsl-EYFP mice crossing Tyr::Cre ERT2 mouse to the ROSA26-lsl-EYFP. Tyr::Cre ERT2 and Tyr::Cre ERT2 ; ROSA26lsl-EYFP mice were treated with 4OHTx at two days and a half, and we assessed the presence of EYFP positive cells in the lung three days later ( Figure 1D). Indeed, EYFP positive cells were found in the lung of Tyr::Cre ERT2 ; ROSA26-lsl-EYFP mice. The cellular origins of human lung adenocarcinoma and adenomas have been addressed during the last years. NSCLCs frequently express markers of Clara cells or alveolar type II pneumocytes (ATII) and it has been suggested that they arise from a bronchioalveolar stem cell population (BASC) [9,16]. To assess the properties of EYFP positive cells, immunohistochemical analyses to detect Clara Cell antigen (CC10) and Surfactant Protein-C (SP-C), a surface marker of ATII pneumocytes, were performed. Staining for CC10 was detected mainly in cells lining the bronchioles and in a few EYFP positive cells ( Figure 1D). In contrast, the majority of EYFP positive cells expressed SP-C ( Figure 1D), suggesting that they have properties of ATII pneumocytes.
We next examined the functional impact of Braf CA transgene and monitored 4OH-tamoxifen-dependent proliferative responses in neonatal lungs of Braf CA and control mice. According to the Ki67 positive cells, BRaf V600E expression increased lung cell proliferation index from 33.51% 62.389% (n = 8 20X fields from 3 different mice) in 10-days-old WT to 42.56% 62.861% (n = 11 20X fields from 3 different mice) in age-matched Braf CA mice (each 20X field quantified contained between 2500-5000 nucleus; P = 0.0345; Figure 1E).
Tyr::Cre ERT2 ; Braf CA/+ Mice Develop Lung Adenomas Serial histopathological examination of the lungs from Tyr::-Cre ERT2 ; Braf CA/+ transgenic mice was conducted to assess the long-term consequences of sustained BRAF V600E expression in the lung. Careful follow-up and characterization of a large colony of Tyr::Cre ERT2 ; Braf CA/+ , Tyr::Cre ERT2 ; Braf CA/CA and control mice showed that only animals on 4OHTx treatment were prone to lung adenoma development. Twenty percent of Tyr::Cre ERT2 ; Braf CA/+ and 33% of Tyr::Cre ERT2 ; Braf CA/CA mice developed papillary adenomas with an average latency of 31.9 weeks (SD 615.1 weeks) and 31.1weeks (SD619.2 weeks) respectively, that were rarely multifocal ( Figure 2A, Table 1, Figure 2B and 2C). In contrast, all Tyr::Cre ERT2 ; Braf CA/+ (n = 65), as well as Tyr::Cre ERT2 ; Braf CA/CA mice (n = 47) of control cohort remained tumor free (Table 1 and Figure 2D and 2E). To investigate the properties of Braf CA induced tumors, we performed immunohistochemical analyses of CC10 and SP-C. Braf CA -induced tumors were largely negative for CC10 where staining was only detected in cells lining the bronchioles ( Figure 2F and 2G). In contrast, the majority of cells within Braf CA -induced tumors expressed SP-C ( Figure 2H and 2I). Furthermore, analysis of the earliest Braf CA -induced lesions revealed them to express SP-C.

Lkb1 Loss Promotes Progression of BRAF V600E -induced Tumors
As previously described, a feature of the Braf CA -induced lung tumors from these animals is that they fail to progress to carcinoma, and instead, exhibited growth arrest and senescencelike features [7,8]. The serine/threonine kinase, LKB1, is a tumor suppressor that has been found mutated in lung cancer concurrently with KRAS mutations [17,18]. Furthermore, LKB1 haploinsufficiency clearly accelerates KRAS-driven lung cancers in mice, where LKB1-deficient tumors demonstrated shorter latency, an expanded histological spectrum (adeno-, squamous and large-cell carcinoma) and more frequent metastasis compared to tumors mutated in Kras lacking p53 or p16 Ink4a /p19 Arf [19]. Therefore, we interrogated whether Lkb1 loss cooperates with oncogenic BRAF V600E in lung tumor development and/or progression. To this end, we crossed Tyr::Cre ERT2 ; Braf CA/+ and Tyr::Cre ERT2 ; Braf CA/CA mouse strains with Lkb1 flox/flox and activated Cre-recombinase at postnatal day 2.5 by topical treatment with 4OHTx. Neonatal inactivation of Lkb1 in Tyr::Cre ERT2 ; Lkb1 flox/+ or Tyr::Cre ERT2 ; Lkb1 flox/flox mice did not have any effect in lung tumor development. However, deletion of one or both alleles of Lkb1 in Tyr::Cre ERT2 ; Braf CA/+ mice increased tumor frequency from 20% to 32% and 26.7% respectively. This increment in the number of animals developing lung tumors was less pronounced in the case of Tyr::Cre ERT2 ; Braf CA/CA mice, where loss of either one or both alleles of Lkb1 slightly increased tumor incidence from 30.7% to 32.4% and 38.6% respectively ( Table 2). Concurrent Braf CA activation and loss of Lkb1 led to pronounced bronchiole epithelial hyperplasia ( Figure 3A and 3B) followed by the development of papillary adenomas associated to small airways ( Figure 3C and 3D). Notably, this mixed papillary and solid adenomas progressed to malignant adenocarcinomas showing occasionally intra-bronchioles tumor growth ( Table 2, Figures 3E-3H). Those adenocarcinomas contained different cell populations. Some cells showed enlarged nuclei displaying prominent nucleoli. There were also atypical cells with nuclear hyperchromasia and contour irregularities, as well as, cells with hyperchromatic fusiform nuclei ( Figure 3I). Adenocarcinomas harbored areas of necrosis and showed evidence of vascular and lymphatic invasion. This is noteworthy since we did not detected adenocarcinoma in any Braf CA mice at least prior to 70 week of age. We compared these Braf CA ; Lkb1 flox/+ adenocarcinomas with K-Ras +/LSLG12Vgeoinduced lung tumors [6]. A significant number of adult CMV-Cre T/+ ; K-Ras +/LSLG12Vgeo mice developed a broad spectrum of multifocal lesions in lungs, ranging from bronchiole-epithelial hyperplasias to large papillary adenomas and adenocarcinomas [6] ( Figures 3J-3M). Interestingly, and regardless of the presence of Lkb1, we never observed the development of mixed solid and mucinous adenomas and adenocarcinomas in Tyr::Cre ERT2 ; Braf CA/+ mice as it occurs in CMV-Cre T/+ ; K-ras +/LSLG12Vgeo mice ( Figure 3N-3P).

Progression of Braf CA/+ ;Lkb1 flox/+ Lesions is Associated to the Loss of Expression of p53, E-Cadherin and SP-C
Unlike Kras +/LSLG12Vgeo -induced lung tumors, Braf CA -driven tumors failed to progress to carcinomas, but exhibited growth arrest and senescence-like features [7,8]. Braf-mutant lung tumor cells grow for approximately 15-20 cell divisions before undergoing senescence-like growth arrest that could be overcome through mutation of p53 or p16 Ink4a /p19 Arf tumor suppressor genes [7,8]. According to Dankort et al. (2007), both, the up-regulation of p16 and the up-regulation of p53 would appear to be important for inducing senescence in tumors [7,8]. This could be a differential  As mentioned before, Braf CA and Kras +/LSLG12Vgeo -induced tumors stained negative for the Clara Cell antigen CC10 regardless the expression of LKB1 (Figure 4). However, even small lesions of Braf CA and Kras +/LSLG12Vgeo -induced adenomas were positive for the surface marker of alveolar type II pneumocytes SP-C (Figure 4). Expression of SP-C was lost in undifferentiated regions of Braf CA/+ ; Lkb1 flox/+ -induced adenocarcinomas as well as Kras +/LSLG12Vgeo advanced tumors (Figure 4). The epithelial marker E-Cadherin was highly expressed in adenomas (Braf CA and Kras +/LSLG12Vgeo ) and its expression was diminished or lost in Braf CA/+ ;Lkb1 flox/+ carcinomas. This characteristic was also observed in Kras +/LSLG12Vgeo carcinomas al-though, in a less pronounced manner ( Figure 4). As expected, Ki67 staining suggested that Kras +/LSLG12Vgeo and Braf CA/+ ; Lkb1 flox/+ -induced adenocarcinomas had increased cell proliferation rates (p,0.0052 for Braf CA/+ ;Lkb1 flox/+ vs. Braf CA/+ ) (Figure 4 and Figure S1A). In agreement with previous data showing that Braf CA -induced adenomas had elevated amounts of p19 ARF [7], Braf CA adenomas showed a significant increased amount of nuclear p53 compared with Kras +/LSLG12Vgeo -adenomas. In contrast, Braf CA/+ ;Lkb1 flox/+ and KRas +/LSLG12Vgeo -induced adenocarcinomas showed very low expression of p53 (Figure 4 and Figure S1B). Interestingly, staining of BRaf CA/+ ;Lkb1 flox/+ samples for LKB1 provided evidence for loss of heterozygosity (LOH). The results also showed spontaneous loss of LKB1 expression in some KRas +/LSLG12Vgeo -induced adenocarcinomas (3 out of 12 tumors) (Figure 4).

Discussion
Deregulation of RAS signaling pathway has been identified as an important event in lung cancer [1,7,8,20]. Activating mutations in BRAF are found in 3% of non small-cell lung cancers [1,2,3], and mouse models have shown that expressing the mutant form of BRAF (BRAF V600E ) in lung epithelium induce the formation of benign lung tumors [7,8]. In this study, we show that conditional expression of BRAF V600E under the control of a tyrosinasepromoter driven transgenic system, induced bronchiole epithelial hyperplasia and adenoma development. These data suggest the existence of a population of lung tumor initiating cells expressing tyrosinase gene in mouse neonate periods. Importantly, loss of only one allele of tumor suppressor Lkb1, frequently mutated in human lung cancer, cooperated with BRAF V600E in tumor development and progression to malignant adenocarcinomas.
We initially observed tyrosinase driven BRAF V600E -induced lung tumor development where Cre-recombinase ERT2 was activated by topical application of 4OHTx in neonate mice (2.5 days old). We and other groups have not observed the development of this tumors when mouse where topical treated with 4OHTx at 6-8 weeks of age (data not shown) [21], suggesting that we were targeting some progenitor cells in immature lungs. Indeed, at this time point we observed Cre-recombinase expressing cells in Tyr::Cre ERT2 ; Braf CA/+ mice. As a result of BRAF V600E expression, we also identified increased amounts of p-ERK1/2 in localized patches at 1.5-2 days after 4OHTx treatment, followed by an increased lung cell proliferation index at six days after treatment. The existence of Cre-recombinase expressing cells was genetically confirmed using the Tyr::Cre ERT2 ;ROSA-lsl-EYFP mouse. These results are in agreement with a previous report describing tyrosinase RNA transcripts in lung tissue [22] and bring attention to the timing at which Cre-recombinase ERT2 produced in animal models harboring Tyr::Cre ERT2 construct should be activated. Whether this cell population represents cells from neural crest cell origin, remains an interesting question that should be investigated in more detail. Nevertheless, these results identify a tyrosinase expressing lung cell population in neonate mice where mutation of BRAF led to adenoma development. In this matter, it has previously been suggested that lung adenomas and adenocarcino-mas arise from a common bronchio-alveolar stem cell population (BASC). These cells express both, Clara Cell antigen (CC10) and Surfactant Protein-C (SP-C), a surface marker of alveolar type II pneumocytes [9,16]. Previous analysis of the earliest BRAF V600Einduced lung lesions revealed them to express SP-C [7]. In agreement with this, our results show that majority of EYFP positive cells were also positive for surface marker of alveolar type II pneumocytes SP-C three days after 4OHTx treatment. Interestingly, in our model and other studies [7], BRAF V600Einduced tumors were also largely CC10 negative and positive for SP-C, suggesting that they have properties of alveolar type II pneumocytes.
Similarly as observed in skin in vivo in benign nevi, the precursor lesion to melanoma [23], BRAF V600E lung tumors failed to progress to carcinoma, and instead, exhibited growth arrest and senescence-like features that could be overcome through concurring mutation of tumor suppressors p53 or p16 Ink4a /p19 ARF , allowing tumors to progress to carcinomas [7]. Several hypotheses have been proposed to explain the differences between Braf V600E and Kras V12 -induced tumors, which progress to carcinomas. These include the up-regulation of p16 Ink4a via MAPK and the upregulation of p53 via PI3K attenuation in senescence-like BRAF V600E tumors. Also the effector cell hypothesis, where BRAF V600E and KRAS V12 mutations would be targeting unique cell types with differences in replicative potential may explain this observation [8]. In agreement with previous data [7], BRAF V600Einduced adenomas showed higher amounts of nuclear p53 compared to Kras V12 -induced adenomas, supporting the senescence-like phenotype of BRAF V600E -induced adenomas. Additionally, while BASC have been postulated as the tumor cell of origin for the KRAS G12D lung tumor model [9,10,16], our results and others [7,8] showed that BRAF V600E target cells and induced tumors were mainly positive for SP-C supporting the effector cell hypothesis.
Several analyses of human tumors have reported that tumor suppressor LKB1 is frequently mutated and inactivated in lung cancer [13,24,25]. LKB1 mutations seem to be more frequent in poorly differentiated adenocarcinomas than in well-differentiated tumors, and mouse models indicate that LKB1 deficiency alone do not cause lung tumors, however synergize with other mutations. In this matter, synergism of Kras activation and Lkb1 deletion has more pronounced effects in tumor development and progression than combination of Kras activation and loss of p53, p16 Ink4a or p19 ARF tumor suppressors. Additionally, in melanoma it has been shown that oncogenic BRAF V600E suppresses some of the functions of LKB1 as an energy metabolism sensor [26], suggesting that activation of oncogenic BRAF V600E and the loss of LKB1 function/s might collaborate in tumorigenesis. Notably, a single-copy inactivation of Lkb1 allowed BRAF V600E -induced adenomas to progress to adenocarcinomas, where some tumors showed evidence for LOH. Interestingly, we also observed the spontaneous loss of Lkb1 expression in Kras +/LSLG12Vgeo carcinomas supporting the cooperation between the activation of MAPK pathway and the loss of tumor suppressor LKB1 in lung cancer ( Figure 4). Finally, these results indicate that LKB1 inactivating mutations constitute a risk factor for tumor progression of BRAF V600E mutated lung adenomas in human cancer patients. It would be interesting to investigate the mutational status of LKB1 in BRAF V600E mutant lung carcinomas. In this regard, it has recently been described that LKB1 inactivation dictates therapeutic response of non-small cell lung cancer to the metabolism drug phenformin [17], suggesting that patients harboring BRAF V600E mutations and LKB1 deficient could benefit from this therapy. In summary, we show evidence for the existence of tyrosinase and SP-C-expressing cells in the lungs of neonate mouse where activation of oncogenic BRAF V600E induces adenoma development. Importantly concomitant activation of BRAF V600E and loss of a single copy of tumor suppressor Lkb1 led to tumor progression. This model serves a system for dissecting the contribution of Lkb1 loss and the activation of MAPK pathway through mutant BRAF in comparison to mutant KRAS. Most importantly, it emphasizes the role of LKB1 in the progression of BRAF V600E mutant lung adenomas. ten fields (206) from three different tumors rose in three different mice Tyr::Cre ERT2 ; Braf CA/+ or Tyr::Cre ERT2 ; Braf CA/+ ;Lkb1 F/+ and six fields (206) from three different tumors rose in Kras +/ LSLG12Vgeo mice (at 6 months and 11 months after KRAS activation) were quantified. p-value was calculated performing Mann-Whitney's test. Bars 500 mm for magnification. (TIF)