Atypical Protein Kinase Cι is Required for Bronchioalveolar Stem Cell Expansion and Lung Tumorigenesis

Abstract

Protein kinase Cι (PKCι) is an oncogene required for maintenance of the transformed phenotype of non-small cell lung cancer (NSCLC) cells. However, the role of PKCι in lung tumor development has not been investigated. To address this question, we established a mouse model in which oncogenic Kras^G12D is activated by Cre-mediated recombination in the lung with or without simultaneous genetic loss of the mouse PKCι gene, Prkci. Genetic loss of Prkci dramatically inhibits Kras-initiated hyperplasia and subsequent lung tumor formation in vivo. This effect correlates with a defect in the ability of Prkci-deficient bronchioalveolar stem cells (BASCs) to undergo Kras-mediated expansion and morphological transformation in vitro and in vivo. Furthermore, the small molecule PKCι inhibitor aurothiomalate inhibits Kras-mediated BASC expansion and lung tumor growth in vivo. Thus, Prkci is required for oncogene-induced expansion and transformation of tumor-initiating lung stem cells. Furthermore, aurothiomalate is an effective anti-tumor agent that targets the tumor-initiating stem cell niche in vivo. These data have important implications for PKCι as a therapeutic target and for the clinical use of aurothiomalate for lung cancer treatment.

INTRODUCTION

Lung cancer is the leading cause of cancer death in the United States (1). NSCLC is the most prevalent type of lung cancer accounting for ~80% of cases (2). Despite recent treatment advances the five year survival rate of NSCLC patients remains only 15%. Thus, there is a dire need to better understand the key molecular events driving lung tumorigenesis that could lead to more effective means of prevention, diagnosis, prognosis and treatment.

Activating mutations in Kras are among the most common molecular changes found in NSCLC tumors (3, 4) and several transgenic mouse models have demonstrated the critical role of oncogenic Kras in lung tumor initiation and maintenance. Transgenic mice carrying a conditional oncogenic Kras^G12D allele activated by Cre-recombinase (LSL-Kras mice) develop epithelial hyperplasias, lung adenomas and adenocarcinomas (5, 6). A second murine lung cancer model (the Kras^LA mouse) is based on sporadic Kras mutation using a “hit-and-run” gene targeting strategy (7). Kras^LA mice carry a latent activation oncogenic Kras^G12D allele activated by spontaneous somatic recombination. Kras^LA mice invariably develop numerous lung tumors pathologically similar to the human disease (7).

Bronchio-alveolar stem cells (BASCs) are the probable cells of origin of Kras-induced lung tumors in these mouse models (8). BASCs function in lung homeostasis and possess characteristics of regional lung stem cells (8). BASCs reside in bronchio-alveolar duct junctions (BADJs), a niche at the boundary between the terminal bronchioles (TBs) and the alveolar space (AS). BASCs express the hematopoietic stem cell marker Sca-1, Clara cell specific protein (CCSP) and surfactant protein C (SPC), but not the hematopoietic and endothelial cell markers CD45 and Pecam (8). BASCs exhibit stem-like self-renewal and the ability to differentiate into Clara cells and alveolar type II (AT2) cells in vitro (8). Moreover, BASCs expand in response to Kras transformation in vitro, and within the BADJ in vivo, eventually giving rise to adenomas and carcinomas in the LSL-Kras mouse model (8).

We recently demonstrated that PKCι is an oncogene in NSCLC (9, 10). PKCι expression is elevated in NSCLC tumors and cell lines, and PKCι is required for the transformed phenotype of NSCLC cells harboring oncogenic Kras mutation (9). PKCι activates a novel oncogenic PKCι-Par6-Rac1 signaling axis that is necessary for the maintenance of the transformed phenotype of human NSCLC cells (11). We recently identified aurothiomalate as a potent and selective inhibitor of oncogenic PKCι signaling (12, 13). Aurothiomalate inhibits NSCLC cell transformation in vitro and tumorigenicity in vivo by disrupting the interaction between PKCι and Par6, thereby blocking the PKCι-Par6-Rac1 signaling axis (12, 13).

Although PKCι is important in the maintenance of the transformed phenotype of NSCLC cells harboring Kras mutation (11, 14), the role of PKCι in Kras-mediated lung tumor development has not been explored. Here we assess the role of the mouse PKCι gene Prkci, in Kras mediated lung tumorigenesis using a novel bi-transgenic mouse model harboring conditional LSL-Kras and Prkci knock out (Prkci^fl) alleles. LSL-Kras/Prkci^fl/fl mice exhibit a significant impairment in lung hyperplasias and lung tumors formation, indicating that Prkci is necessary for early events in Kras-mediated lung tumorigenesis. The effect of Prkci deficiency on tumorigenesis correlates with a defect in Kras-mediated BASC expansion in the absence of Prkci. Furthermore, the PKCι inhibitor aurothiomalate inhibits Kras-mediated BASC expansion and transformation in vitro and in vivo. Our results demonstrate that Prkci plays a requisite role in BASC transformation and identify aurothiomalate as a promising therapeutic agent that can target the lung cancer stem cell niche in vivo.

MATERIALS AND METHODS

Generation and AdCre treatment of LSL-Kras/Prkci^fl/fl mice

LSL-Kras mice (B6.129), generated as previously described (5), were mated to mice harboring a floxed Prkci allele (Prkci^fl mice) to generate bi-transgenic LSL-Kras/Prkci^fl/fl mice. Non-transgenic littermates served as controls in all experiments. Mice (6–8 weeks of age) were administered AdCre (1.5 × 10⁸ IFU/mL) by intratracheal instillation in two 50 μL aliquots as described previously (15). All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of Mayo Clinic.

Histology and immunohistochemistry

Mice were sacrificed, exsanguinated and the lungs perfused with 10% buffered formalin through the right ventricle. The trachea was intubated and instilled with an additional 3 mL of 10% buffered formalin. The lungs were removed intact and fixed overnight in 10% buffered formalin. For histological analysis, lungs were embedded in paraffin, serially sectioned (5 μm) and stained with hematoxylin and eosin. Immunohistochemical analysis was performed as described (10). Mouse PKCι was detected using a PKCι antibody (Santa Cruz Biotechnology) and visualized using the Envision Plus Dual Labeled Polymer Kit following the manufacturer's instructions (DAKO). In some experiments, the antibody was incubated overnight at 4°C with a 200-fold molar excess of PKCι peptide (Santa Cruz Biotechnology) prior to use in immunohistochemistry to confirm antibody specificity. Slide images were captured and analyzed using the ScanScope scanner and ImageScope software (Aperio Technologies).

BASC isolation and culture

Lung epithelial cells were isolated as described (16). Red blood cells were lysed in RBC lysis buffer (StemCell Technologies) and BASCs isolated using the EasySep® immunomagnetic cell selection procedure (Stem Cell Technologies). Briefly, CD45^pos Pecam^pos cells were selected out using primary biotinylated antibodies anti-CD45 (BD Pharmingen) and anti-Pecam (BD Pharmingen). CD45^neg Pecam^neg cells were incubated for 15 minutes at room temperature with FITC-conjugated anti-CD34 (BD Pharmingen) and Sca1-PE labeling reagent (Stem Cell Technologies). Sca1^pos CD34^pos CD45^neg Pecam^neg BASCs were selected using the EasySep® FITC and PE selection kits. BASCs were resuspended in BEGM (Lonza, without hydrocortisone) containing 10ng/mL keratinocyte growth factor (PeproTech) and 5% charcoal-stripped fetal bovine serum (FBS) and plated at equal densities on Matrigel-coated (BD Biosciences) tissue culture wells. Cultures were infected with AdCre (7.5 × 10⁷ IFU/mL) the day after plating and medium was changed every 2 days thereafter. Brightfield images of BASC colonies were captured on an Olympus IX71 inverted microscope and colony size was determined using Image-Pro Plus 6.3 software (Media Cybernetics).

Immunofluorescence of BASCs in Matrigel culture was performed as described (17) using the following antibodies: rabbit anti-CCSP (Upstate), goat anti-SP-C (Santa Cruz Biotechnology), rabbit anti-Ki67 (Abcam), Alexa Fluor 488 donkey anti-rabbit IgG, Alexa Fluor 568 donkey anti-goat IgG, and Alexa Fluor 594 goat anti-rabbit IgG (Molecular Probes). Samples were mounted using ProLong Gold anti-fade reagent with DAPI (Molecular Probes) and observed on a Zeiss 510 LSM confocal laser microscope. Images were analyzed with Adobe Photoshop software (Adobe Systems).

BASC colonies were released from Matrigel culture using BD cell recovery solution (BD Biosciences) and total RNA subjected to quantitative PCR as described (18). RNA quantity and integrity were measured using a NanoDrop® ND-1000 spectrophotometer.

Kras and Prkci allele status

Kras allele status in AdCre-treated BASC cultures and adenomas microdissected from LSL-Kras/Prkci^fl/fl mice was determined by PCR as described (5, 19). Prkci^fl allele status was determined using the following PCR primers: F1 = 5'-CACCAACGGGTTTGCTATCT-3', R1 = 5'-CCAGCAAGACAAAACACCAA-3', R2 = 5'-ATTACAGCAGGGCAAACTGC-3'.

Treatment of Kras^LA2 mice with aurothiomalate

Three-week-old Kras^LA2 mice were given daily intraperitoneal injections of aurothiomalate (Myochrysine®; Taylor Pharmaceuticals) at 60 mg/kg/day or vehicle control (0.9% sodium chloride solution, Sigma-Aldrich) for the indicated time period. Tumors were analyzed as described above.

Detection and analysis of BASCs in mouse lung tissues

BASCs were detected in formalin fixed paraffin embedded (FFPE) mouse lung tissues as described previously (20). Lung tissue sections were stained using goat anti-SPC (Santa Cruz Biotechnology) and rabbit anti-CCSP (Upstate) antibodies diluted in 1% BSA (Sigma-Aldrich) at 4°C overnight. The slides were washed thoroughly with PBS and then incubated with Alexa Fluor 488 donkey anti-rabbit IgG and Alexa Fluor 568 donkey anti-goat IgG secondary antibodies (Molecular Probes) diluted in 1% BSA for 1 hour at room temperature. After washing with PBS, the samples were mounted using ProLong Gold antifade reagent with DAPI. Immunofluorescence images were captured on an Olympus BX51 fluorescent microscope using DPController and DPManager software (Olympus), and processed in Adobe Photoshop.

Statistical analysis

The Cochrane-Armitage test was used to compare BASC number and distribution in vivo using StatsDirect 2.6.1. The Student's t test and one-way ANOVA statistical analyses were performed using SigmaStat 3.5. A P value of less than 0.05 was considered statistically significant.

RESULTS

Prkci is required for Kras-mediated lung tumorigenesis

To determine whether PKCι is involved in lung tumor formation, we employed the LSL-Kras mouse model in which lung tumors are initiated by AdCre-mediated activation of a conditional oncogenic Kras allele (5). Since PKCι is overexpressed in primary human NSCLC tumors (10), we assessed whether mouse PKCι protein is also highly expressed in lung tumors from LSL-Kras mice. Immunohistochemical analysis revealed light staining for mouse PKCι in normal lung epithelial cells with more intense staining in lung tumor cells from LSL-Kras mice (Figure 1A, right panel). Thus, mouse PKCι is highly expressed in Kras-initiated lung tumors in mice.

Figure 1. Prkci is necessary for Kras-mediated lung tumorigenesis.

LSL-Kras and LSL-Kras/Prkci^fl/fl mice were infected with AdCre, harvested and processed as described in Materials and Methods. (A) Staining of a LSL-Kras lung tumor for mouse PKCι protein. (B) Histopathology of LSL-Kras and LSL-Kras/Prkci^fl/fl mice. upper left: multiple lesions are evident on the surface of LSL-Kras mouse lungs (arrows). upper middle: LSL-Kras mice develop numerous focal adenomas (arrows) upper right: LSL-Kras adenomas often surround bronchioles. lower left: LSL-Kras/Prkci^fl/fl mice exhibit a smooth lung surface devoid of visible lesions. lower middle: LSL-Kras/Prkci^fl/fl lungs exhibit fewer and smaller lesions than LSL-Kras mice. lower right: LSL-Kras/Prkci^fl/fl lung exhibit largely normal lung morphology. (C) LSL-Kras/Prkci^fl/fl mice exhibit reduced lung hyperplasias. Mean +/− SE (n= 6); *P < 0.05. (D) LSL-Kras/Prkci^fl/fl mice exhibit reduced lung tumor formation. Mean +/− SE (n= 12); *P = 0.007.

We have shown that PKCι is required for transformed growth and invasion of human NSCLC cells in vitro and tumorigenicity in vivo (9, 10). To assess the role of PKCι in oncogenic Kras-initiated lung tumorigenesis, we crossed mice harboring a conditional knockout Prkci allele (Prkci^fl) with LSL-Kras mice to generate bi-transgenic LSL-Kras/Prkci^fl/fl mice. In LSL-Kras/Prkci^fl/fl mice, intratracheal instillation of AdCre leads to simultaneous Cre-mediated activation of the oncogenic Kras allele and inactivation of the Prkci^fl/fl alleles in the same target cells within the lung. Twelve weeks following AdCre administration, LSL-Kras and LSL-Kras/Prkci^fl/fl mice were assessed for lung hyperplasias and lung tumors (Figure 1B). As expected, LSL-Kras mice developed numerous tumors visible on the surface of the lung (Figure 1B, upper left panel, arrows). In contrast, LSL-Kras/Prkci^fl/fl mice exhibited few visible tumors (Figure 1B, lower left panel). Pathological examination revealed that LSL-Kras mice develop multiple focal adenomas often surrounding bronchioles (Figure 1B, upper middle and right panels, arrows) whereas LSL-Kras/Prkci^fl/fl mice exhibit largely normal lung morphology and fewer lesions (Figure 1B, lower middle and right panels). Analysis also revealed a significant reduction in both hyperplasias (Figure 1C) and adenomas (Figure 1D) in LSL-Kras/Prkci^fl/fl mice when compared to LSL-Kras mice. Thus, genetic deletion of Prkci inhibits hyperplasia and subsequent tumor formation, indicating a role for Prkci in an early step of Kras-mediated lung tumorigenesis.

Since LSL-Kras/Prkci^fl/fl mice formed some lung tumors, albeit in significantly reduced numbers, we determined the status of the Kras and Prkci alleles in tumors formed in these mice. Genomic DNA from tumors microdissected from LSL-Kras/Prkci^fl/fl mice was analyzed for Kras and Prkci allele status by PCR (Figure 2A and B). As expected, the LSL-Kras allele had undergone Cre-mediated recombination (Figure 2A, top panel) with concomitant loss of the unrecombined Kras allele (Figure 2A, bottom panel) in all of the tumors analyzed. In contrast, each tumor contained a recombined (deleted) Prkci allele (Figure 2B, top panel), and an unrecombined Prkci allele (Figure 2B, bottom panel), indicating incomplete deletion of the conditional Prkci^fl alleles. The presence of unrecombined Prkci was not due to contamination of the tumor samples by normal lung as indicated by the absence of the unrecombined LSL-Kras allele in the same tumor samples (Figure 2A, bottom panel). Furthermore, immunohistochemistry confirmed expression of mouse PKCι protein in these tumors (Figure 2C). Our data indicate that lung tumors in LSL-Kras/Prkci^fl/fl mice arise due to incomplete recombination of the conditional Prkci^fl alleles. Thus, Prkci is critical for Kras-induced lung tumorigenesis, and only Kras-transformed cells that escape complete inactivation of both Prkci^fl alleles are able to progress in tumorigenesis.

Figure 2. Lung tumors in LSL-Kras/Prkci^fl/fl mice arise due to incomplete recombination of the Prkci^fl alleles.

Genomic DNA from microdissected LSL-Kras/Prkci^fl/fl tumors was analyzed for Kras (A) and Prkci (B) allele status. Each tumor expresses the recombined (A, top) but not the unrecombined Kras allele (A, bottom). Normal lung epithelium (N) and tumor (T) DNA from a Kras mouse serve as controls for Kras allele recombination. All LSL-Kras/Prkci^fl/fl tumors analyzed contained a recombined Prkci allele (B, top panel) as well as an unrecombined Prkci allele (B, bottom panel), indicating incomplete deletion of the conditional Prkci^fl alleles. (R) recombined and (U) unrecombined Prkci allele controls. (C) LSL-Kras/Prkci^fl/fl tumors express mouse PKCι protein. Adjacent sections were stained with H&E (left), PKCι antibody (middle), or with PKCι antibody in the presence of PKCι peptide to verify specificity of staining (right).

Prkci is required for Kras-mediated expansion of the bronchio-alveolar stem cell (BASC) niche in vivo

Our data indicate that Prkci is required for a very early event(s) in Kras-mediated oncogenesis in vivo. Kras mediated lung tumorigenesis in the LSL-Kras mouse model is thought to be initiated by clonal expansion of BASCs, the putative cell of origin of Kras-initiated lung tumorigenesis (8). Therefore, we next assessed whether Prkci-deficiency affects the expansion of BASCs induced by Kras in vivo. BASCs are identified in tissue sections at the BADJ by dual immunofluorescence for SPC and CCSP (Figure 3A). AdCre-treated LSL-Kras mice exhibit frequent clusters of SPC-, CCSP-double-positive BASCs at BADJs indicative of BASC expansion (Figure 3A, upper panels). In contrast, LSL-Kras/Prkci^fl/fl mice exhibit significantly fewer BASCs at the BADJs (Figure 3A, lower panels). Quantitative analysis of sections from AdCre-treated non-transgenic (Ntg), LSL-Kras and LSL-Kras/Prkci^fl/fl mice revealed a significant increase in the percentage of terminal bronchioles (TBs) containing one or more BASCs, and in the average number of BASCs/TB in LSL-Kras mice when compared to Ntg mice (Figure 3B). LSL-Kras/Prkci^fl/fl mice exhibited a significant inhibition in Kras-mediated BASC expansion, such that the number and distribution of BASCs in these mice was indistinguishable from that of Ntg mice (Figure 3B). Furthermore, AdCre-treated Prkci^fl/fl mice exhibited no difference in BASC number or distribution compared to Ntg mice, indicating that Prkci deficiency does not overtly affect the BASC niche in the absence of oncogenic Kras (Figure 3B). Thus Prkci is necessary for Kras-mediated BASC expansion in vivo, linking the effect of Prkci-deficiency on BASC expansion with lung tumorigenesis. Furthermore, the inhibitory effect of Prkci deficiency on lung tumorigenesis is not due to a decrease in the number of BASCs in LSL-Kras/Prkci^fl/fl mice.

Figure 3. Prkci is necessary for Kras-mediated expansion of BASCs in vivo.

LSL-Kras and LSL-Kras/Prkci^fl/fl mice were harvested 12 weeks after intratracheal instillation of AdCre. Lung tissue sections were subjected to immunofluorescence as described in Materials and Methods. (A) Expansion of CCSP (green) and SPC (red) double positive BASCs (outlined) is evident within the terminal bronchioles (TB) of LSL-Kras mice (upper panels). Clara cells within the bronchioles stain positive for CCSP but negative for SPC. AS = alveolar space. BASC expansion is not obsevrved in LSL-Kras/Prkci^fl/fl mice (lower panels). (B) Quantitative analysis of terminal bronchioles for BASCs in lungs from Ntg, Prkci^fl/fl, LSL-Kras and LSL-Kras/Prkci^fl/fl mice. Results are expressed as % terminal bronchioles +/− SE. Statistical analysis revealed a significant increase in BASC number and BASC/TB in LSL-Kras mice compared to Ntg, Prkci^fl/fl or LSL-Kras/Prkci^fl/fl mice (p<0.0001). No statistically significant difference in BASC number or distribution was observed between Ntg, Prkci^fl/fl and LSL-Kras/Prkci^fl/fl mice.

Oncogenic Kras induces expression of PKCι and Prkci-dependent proliferative signaling in BASCs in vitro

To further explore the role of Prkci in BASC expansion, we isolated and established BASCs from Ntg, LSL-Kras and LSL-Kras/Prkci^fl/fl mice in primary culture using established procedures (8). Isolated lung epithelial cell populations consisted of a mixture of SPC-single positive AT2 cells (Figure 4A, left panel, AT), CCSP-single positive Clara cells (Figure 4A, left panel, arrows), and SPC-, CCSP-double positive BASCs (Figure 4A, left panel, arrowheads). This BASC population was enriched based on positive staining for the stem cell markers Sca1 and CD34. Consistent with previous reports, >95% of the cells remaining after this selection stained double positive for the BASC markers SPC and CCSP (Figure 4A, right panel). Purified BASCs from Ntg, LSL-Kras and LSL-Kras/Prkci^fl/fl mice were treated with AdCre, and efficient recombination of the LSL-Kras and Prkci^fl alleles was confirmed by PCR (Figure 4B). Consistent with our finding that mouse PKCι protein expression is elevated in lung tumors from LSL-Kras mice, AdCre treatment of BASCs from LSL-Kras mice induced a significant increase in PKCι mRNA abundance, whereas BASCs from LSL-Kras/Prkci^fl/fl mice, as expected, expressed no detectable PKCι mRNA (Figure 4C). We previously identified matrix metalloproteinase 10 (MMP10) as a transcriptional target of the oncogenic PKCι-Par6-Rac1 signaling axis (11), and MMP10 has also been identified as a transcriptional target of oncogenic Kras (21, 22). Therefore, we assessed MMP10 expression in BASCs from LSL-Kras and LSL-Kras/Prkci^fl/fl mice by QPCR. MMP10 mRNA was induced in BASCs from LSL-Kras mice but not LSL-Kras/Prkci^fl/fl mice after AdCre treatment (Figure 4D). Thus oncogenic Kras activates oncogenic PKCι signaling in BASCs, and this signaling is inhibited in BASCs from LSL-Kras/Prkci^fl/fl mice.

Figure 4. Kras induces PKCιexpression and activates oncogenic Prkci-mediated signaling in BASCs in vitro.

(A) Immunofluroescence staining of isolated lung epithelial cells (left) and purified BASCs (right) from LSL-Kras mice for CCSP (green), SPC (red), and DAPI (blue). SPC-positive AT2 cells (AT), CCSP-positive Clara cells (arrows) and SPC-, CCSP-double positive BASCs (arrowheads). (B) LSL-Kras and Prkci^fl allele status in BASCs isolated from Ntg, LSL-Kras and LSL-Kras/Prkci^fl/fl mice in the presence and absence of AdCre. The conditional Kras and Prkci alleles are efficiently recombined in the presence of AdCre. (C) Kras induces PKCι mRNA expression in BASCs from LSL-Kras but not LSL-Kras/Prkci^fl/fl mice. (D) Kras induces MMP10 mRNA in BASCs from LSL-Kras but not LSL-Kras/Prkci^fl/fl mice. Data in C and D represent the mean +/− SE of triplicate determinations. Results are representative of four independent experiments with similar results. * indicates a statistically significant difference from LSL-Kras in the absence of Cre, p<0.05; ** indicates a statistically significant difference from LSL-Kras, p<0.05.

Oncogenic Kras-mediated BASC expansion and morphological transformation requires Prkci

Since BASCs expand and initiate Kras-mediated tumors in vivo (8, 23), we assessed BASCs from LSL-Kras and LSL-Kras/Prkci^fl/fl mice for proliferative expansion and morphological transformation after induction of oncogenic Kras in vitro (Figure 5). BASCs from Ntg mice treated with AdCre form colonies of organized spherical structures when plated in three dimensional Matrigel culture (Figure 5A, left panel). In contrast, BASCs from LSL-Kras mice grow as large amorphic, disorganized colonies (Figure 5A, middle panel). Interestingly, BASCs from LSL-Kras/Prkci^fl/fl mice form spherical colonies similar in size and morphology to those from Ntg mice (Figure 5A, right panel). Quantitative analysis revealed that BASCs from LSL-Kras mice form structures of larger average diameter (Figure 5B) and more amorphic shape (Figure 5C) than BASCs from either Ntg or LSL-Kras/Prkci^fl/fl mice. Thus the change in colony size and morphology observed in LSL-Kras BASCs requires both Kras and Prkci. Immunofluorescence confocal microscopy revealed that BASCs from Ntg and LSL-Kras/Prkci^fl/fl mice form hollow spheres consisting of a single layer of epithelial cells with a low proliferative index; in contrast, BASCs from LSL-Kras mice form solid masses of highly proliferative cells (Figure 5D, left panel). Quantitative analysis demonstrated a significant increase in proliferation of LSL-Kras BASCs (% of cells staining positive for Ki67) when compared to Ntg BASCs and that this increase was not seen in LSL-Kras/Prkci^fl/fl BASCs (Figure 5D, right panel). Thus, oncogenic Kras induces Prkci-dependent proliferative expansion and morphological changes associated with cellular transformation of BASCs in vitro.

Figure 5. Kras induces Prkci-dependent BASC expansion and morphological transformation in vitro.

BASCs from Ntg, LSL-Kras and LSL-Kras/Prkci^fl/fl mice were treated with AdCre and plated in three dimensional Matrigel culture. (A) Photomicrographs of cultures after four days. BASC colonies were analyzed for diameter (B) and morphology (C). Data in (B) represent the mean +/− SE and are presented as % Ntg control. Data in (C) represent the % of colonies with spherical (filled bar) or amorphic (open bar) morphology. *indicates a statistically significant difference from Ntg; ** indicates a statistically significant difference from LSL-Kras; p<0.0001. There was no statistically significant difference in colony diameter or morphology between Ntg and LSL-Kras/Prkci^fl/fl colonies. (D) Confocal microscopy of BASC colonies stained for DAPI and Ki67 (left). Quantitative analysis of Ki67 staining (right). Data represent the mean +/−SE expressed as the % of Ki67 positive cells; *indicates a statistically significant difference from Ntg; ** indicates a statistically significant difference from LSL-Kras; p<0.05.

BASCs exhibit self-renewal and the ability to differentiate into bronchiolar Clara cells and AT2 cells (8). Therefore, we assessed whether the increase in proliferation observed in LSL-Kras BASC cultures in vitro was due to expansion of undifferentiated BASCs, or alternatively induction of lineage-specific differentiation of BASC into single positive Clara and/or AT2 cells. Confocal immunofluorescence microscopy revealed that BASC colonies from Ntg, LSL-Kras and LSL-Kras/Prkci^fl/fl mice are made up pre-dominantly of SPC-,CCSP-double positive staining BASCs (Supplemental Figure 1), indicating that Kras induces Prkci-dependent BASC expansion.

Aurothiomalate blocks Kras-mediated, Prkci-dependent BASC expansion and morphological transformation in vitro

We recently identified aurothiomalate as a highly selective small molecule inhibitor of oncogenic PKCι/λ signaling that exhibits potent anti-tumor activity against human NSCLC cells (12, 13). Since Prkci is necessary for Kras-mediated BASC expansion and morphological transformation in vivo and in vitro, we assessed whether aurothiomalate inhibits the effects of Kras on BASCs. BASCs from LSL-Kras mice were transduced with AdCre and plated in three dimensional Matrigel culture in the presence of 20 μM aurothiomalate or diluent (saline) (Figure 6A, left and middle panels). As expected, saline-treated BASCs form large amorphic colonies; in contrast aurothiomalate-treated cells form small spherical colonies similar to those formed by BASCs from Ntg and LSL-Kras/Prkci^fl/fl mice (Figure 6A, right panel). Analysis revealed that saline-treated BASCs form significantly larger colonies than aurothiomalate-treated BASCs (Figure 6B, left panel), that is associated with an increased proliferative index compared to aurothiomalate-treated BASCs (Figure 6B, right panel). Thus, aurothiomalate, like genetic loss of Prkci, blocks Kras-mediated proliferative expansion and morphological transformation of BASCs.

Figure 6. Aurothiomalate inhibits Kras-mediated BASC expansion and morphological transformation in vitro and tumor growth in vivo.

(A) Morphology of BASCs from LSL-Kras mice in the presence of 20 μM aurothiomalate (middle) or saline (left). Analysis of BASC colony morphology in the presence and absence of aurothiomalate (right). * indicates a statistically significant difference from BASCs cultured in the absence of aurothiomalate; p<0.001. (B) Colony diameter (left) and Ki67 staining (right) of BASCs in the presence and absence of aurothiomalate. Data represent the mean +/−SE; *indicates a statistically significant difference from BASCs in the absence of aurothiomalate; p=0.01. (C) Aurothiomalate blocks Kras-mediated BASC expansion in vivo. Ntg and Kras^LA2 mice were treated with aurothiomalate or saline for 3 weeks and lungs were analyzed for BASC number and distribution as described in Figure 3. Statistical analysis revealed a significant increase in BASC number and in the mean number of BASC/TB in saline-treated Kras^LA2 mice when compared to aurothiomalate-treated Kras^LA2 mice, Ntg mice treated with aurothiomalate, or Ntg mice treated with saline (p<0.0001). No statistically significant differences in BASC number or distribution were observed between saline-treated Ntg, aurothiomalate-treated Ntg and aurothiomalate-treated Kras^LA2 mice. (D) Aurothiomalate inhibits lung tumor growth in Kras^LA2 mice. Data represent the mean +/− SE fold-change in average tumor size after six weeks in the presence of aurothiomalate or saline. *indicates a statistically significant difference from saline-treated Kras^LA2 mice; p=0.003.

Aurothiomalate inhibits Kras-mediated BASC expansion and lung tumorigenesis in vivo

We next assessed the effect of aurothiomalate on BASC expansion and tumor formation in Kras^LA2 mice harboring a latent oncogenic Kras allele activated by spontaneous recombination in vivo (7). Ntg and Kras^LA2 mice were treated with either aurothiomalate (60 mg/kg/day) or diluent (saline) for three weeks at which time mice were sacrificed and assessed for BASC number and distribution as described above. Aurothiomalate has no demonstrable effect on BASC number or distribution in Ntg mice (Figure 6C) consistent with our finding that Prkci-deficiency does not overtly affect BASCs in the absence of oncogenic Kras in vivo. In contrast, saline-treated Kras^LA2 mice exhibit a significant increase in BASCs/TB that is inhibited by aurothiomalate (Figure 6C). Consistent with the inhibitory effect of aurothiomalate on BASC expansion, aurothiomalate-treated Kras^LA2 mice exhibited a decrease in tumor growth when compared to saline-treated mice after a six week course of aurothiomalate treatment (Figure 6D). Therefore, like genetic loss of Prkci, aurothiomalate blocks Kras-mediated BASC expansion and inhibits subsequent lung tumor growth in vivo.

DISCUSSION

The PKCι oncogene plays a requisite role in the maintenance of the transformed phenotype of NSCLC cells in vitro and in vivo (10). However, the present study is the first to assess the role of PKCι in lung tumor formation. Our data firmly establish that genetic disruption of Prkci has a profound inhibitory effect on oncogenic Kras-mediated lung tumor development. Although Prkci-deficient mice still develop lung hyperplasias and lung tumors in response to oncogenic Kras, albeit at a dramatically reduced incidence, molecular analysis of these tumors demonstrates that these lesions invariably retain an unrecombined Prkci allele and express mouse PKCι protein. These data argue that Prkci is absolutely required for a very early step in Kras-mediated lung tumorigenesis, such that only the occasional Kras-transformed cell that escapes complete recombination of both Prkci^fl alleles (and retains PKCι expression) is competent to progress in tumorigenesis.

The inhibitory effect of Prkci deletion on tumorigenesis is reflected in a severe defect in the ability of Prkci-deficient BASCs to expand and undergo Kras-mediated transformation in vitro. Thus Prkci is necessary for the first identifiable step in Kras-mediated tumorigenesis involving expansion and transformation of the putative lung cancer stem cell in the mouse. Our data provide important new insight into a key pathway involved in BASC expansion in response to an oncogenic stimulus; namely the oncogenic PKCι signaling axis. Our data place Prkci among a small cadre of genes known to play a critical role in oncogene-induced BASC expansion, including the polycomb group member Bmi1 (23), the p38 MAPK gene (24), the cyclin-dependent kinase inhibitor p27 (25) and phosphatidylinositol 3-kinase (PI3K) (20).

By virtue of its role in BASC expansion, PKCι is an attractive target for therapeutic intervention in lung cancer. Aurothiomalate, a highly selective inhibitor of oncogenic PKCι/λ signaling (12, 13), potently blocks Kras-mediated expansion and morphological transformation of BASCs in vitro and in vivo, and inhibits lung tumor growth in vivo. Our data indicate that aurothiomalate suppresses oncogene-mediated tumor induction, at least in part, through inhibition of the expansion of lung cancer stem cells. These data have important implications for the use of aurothiomalate in cancer treatment. A population of cells within human lung tumors has been identified that exhibits properties of lung cancer stem cells similar to those exhibited by BASCs (26, 27). These lung cancer stem cells are capable of self-renewal, serial passage as tumors in immunocompromised mice, and differentiation into highly proliferative tumor cells (26, 27). Based on these properties, lung cancer stem cells are thought to be responsible for initiation and long-term maintenance of human lung tumors, much as BASCs are in the mouse lung. Given their central role in tumorigenesis, lung cancer stem cells must be therapeutically targeted to achieve lasting anti-tumor responses. Unfortunately, lung cancer stem cells exhibit intrinsic resistance to commonly used cytotoxic agents (26, 27). Our present results demonstrate that aurothiomalate exhibits potent anti-proliferative activity toward the tumor stem cell niche in a relevant pre-clinical lung cancer model. Future studies will be required to assess whether aurothiomalate has similar anti-proliferative effects on human lung cancer stem cells isolated from primary human lung tumors.