Introduction

The superfamily of PI3K (phosphoinositide 3-kinase) plays a fundamental role in homeostasis by regulating several key cellular activities such as proliferation, metabolism, and survival [1, 2]. To the same extent, deregulation of PI3K pathway due to genetic abnormalities such as mutation or amplification of components of the pathway (particularly PIK3IA) is frequently observed in several human cancers. Among four classes of PI3K family, class IA consists of a regulatory subunit (members of p85) and a catalytic subunit (p110α) known to be more frequently involved in cancer occurrence and progression [3, 4].

Kras mutation is one of the major drivers of tumor progression in several cancer types including NSCLC (non-small cell lung cancer) [5], pancreatic carcinomas [6], and CRCs (colorectal carcinoma) [7]. Therefore, KrasG12D-LSL genetically engineered mouse model (GEMM) may be one of the most relevant models to study tumor growth and progression in NSCLC and testing the efficacy of anti-cancer compounds [8]. Recent investigations suggest that KrasG12D-LSL GEMM may closely mimic the clinical trial outcome in cancer patients [9].

Activated Kras also has the ability to bind to p110α resulting in the activation of PI3K pathway. mTORC1 & 2 (mammalian target of rapamycin containing protein complex) are also involved in the activation of PI3K-AKT that in turn results in the activation of downstream cellular functions [10, 11]. Although Ras has the ability to activate PI3K pathway, inhibition of PI3K pathway in established tumors in Kras-mutated lung model does not completely inhibit tumor growth indicating role of other signaling pathways such as MAPK in the growth of NSCLCs [12]. Inhibition of PI3K/mTOR pathway may result in the activation of MAP kinase pathways through compensatory mechanisms. Similarly, treatment of Kras-mutant cancer cell lines with MEK inhibitor may activate the PI3K pathway leading to lack of response to the treatment [13]. Given the role of MAPK and PI3K in the progression of a variety of tumor types including NSCLC, pancreatic ductal carcinoma, and CRC, several inhibitors of the above pathways are in different stages of clinical trials [1, 14, 15], particularly dual inhibitors of PI3K and mTOR pathways as they appear to be more effective than selective inhibitors to block mTOR pathway when it is activated independent of PI3K [14, 16]. Recent investigations identified PF-04691502 (hereafter PF-502) as a potent dual inhibitor of PI3K/mTOR [17] and PD-0325901 (hereafter PD-901) as an inhibitor of MEK pathway [18]. A recent report showed anti-tumor activity of PF-502 in several preclinical models [19], while PD-901 has shown potent tumor growth inhibition in papillary thyroid carcinoma cells [20]. Very recent report in a GEMM of ovarian cancer (KrasG12D-LSLPten−/−) showed that combination of PD-502 and PD-901 significantly inhibited tumor growth and increased survival compared with either monotherapy [21].

Despite potent anti-tumor activity, PD-901 administration in patients at 15 mg or higher has been associated with several adverse effects including visual disturbances secondary to retinal vein occlusion (RVO) [22–24]. In an effort to alleviate potential toxicity without compromising anti-tumor efficacy, we tested efficacy of combination therapy using PD-901 at sub-MTD dose (1.5 mg/kg) in a human xenograft cell line (NCI-H460) and also in KrasG12D-LSL mice. Our investigations indicate that PD-901 is still a potent inhibitor of MAPK pathway even at doses below MTD as single agent and particularly in combination with PF-502.

Materials and methods

Tumor implantation

All animal experimentation was approved by the Pfizer La Jolla Institutional Animal Care and Use Committee (IACUC). Human H460-Luc (luciferase) expressing cells were grown in RPMI1640 supplemented with 10 % FCS. Cells were injected iv (intravenous via tail vein) in Scid/beige mice (1 × 106 in 200 μl PBS). Mice were randomized at 7 days after implantation when tumors were established in the lungs and were treated with Vehicle, PF-502 (7.5 mg/kg; SID, MTD: 10 mg/kg [19]), PD-901 (1.5 mg/kg; SID, MTD: 25 mg/kg [24]) and combination of both compounds. For Luc imaging, mice were injected intraperitoneally with luciferin (at randomization stage 15 mg/ml and for other time points 3 mg/ml, Promega, WI) and were anesthetized with isoflurane. Anesthetized animals were imaged in a Xenogen IVIS-100 instrument (Caliper Biosciences, MA, USA). Luc signals were normalized to 1,000–30,000 counts/s, analyzed and quantified using Living Image 4.0 Software (Caliper, MA).

Tumor development in KrasG12D-LSL GEMMs

KrasG12D-LSL heterozygous mice were obtained from Jackson Laboratories (Jax West, CA) at approximately 3–4 weeks of age. Lung tumors were generated in KrasG12D-LSL mice, using a previously published protocol developed by DuPage et al. [25]. Adenovirus expressing Cre recombinase (Adeno-Cre) was purchased from the University of Iowa Gene Transfer Vector Core. Viral titration was determined by Adenoviral Titration Kits (Clontech, CA) using the manufacturer’s protocol. Mice were administered with 2.5 × 107 infectious units (IU) of the adenovirus by inhalation. Prior to inhalation, Adeno-Cre (2.5 × 107 IU) was prepared in 50 μl media (MEM, Gibco BRL; Life Sciences, CA) containing 10 mM CaCl2 followed by incubation at room temperature (RT) for 20 min. KrasG12D-LSL mice were anesthetized using Ketamine and Xylazine and the Adeno-Cre preparation was administered intranasally. Using the above procedure, we were able to consistently and reproducibly generate tumors in KrasG12D-LSL mice. To ensure tumor formation and progression, KrasG12D-LSL mice (n = 1–3) were euthanized at several time points (4, 8, 12 and 16 weeks post-inhalation) and lung tissues were fixed in formalin. Sections of lungs were stained with H&E (Hematoxylin and Eosin) using standard protocols provided by Department of Drug Safety Research and Development (DSRD). In a series of studies, we observed that using our optimized method, tumors were approximately distributed evenly in the lung. In addition, using serial sectioning, it was evident that the distribution of tumors was consistent between the surface and the parenchyma of the lung tissue, providing us with the confidence to analyze a single lobe of lung possibly as a representative of the entire organ.

To score lesions (hyperplastic, benign adenoma, and adenocarcinoma) in each lung, all the H&E slides were analyzed by a certified pathologist who was provided detail of the study design and the treatments. Hyperplastic lesions were defined by increased density in alveolar epithelial cells with uniform size and shape and no evident mitotic activity. Adenoma lesions were clusters of alveolar epithelial cells that were monomorphic, hypertrophic and were arranged in multiple layers with papillary extensions. Adenocarcinoma lesions were defined by tumor epithelial cells (generally larger than hyperplastic/adenoma cells) in different size and shape that were clustered in papillary forms or dense sheets.

Therapeutic compounds

At 8 weeks post-inhalation, lung tissues were isolated from the KrasG12D-LSL recipient mice to ensure tumor formation. Treatment with anti-cancer agents was initiated at 14 weeks after inhalation with Adeno-Cre when lung lesions are at advanced stage (adenocarcinoma) of tumorigenesis in KrasG12D-LSL mice (3, 5). Tumor-bearing mice were treated with anti-cancer compounds including Pfizer selective mTOR inhibitor hereafter PSMI; 80 mg/kg; SID); PF-502 (7.5 mg/kg; SID); PD-901 (1.5 mg/kg; SID); combination of PF-502 and PD-901 at the respected dose; two standard of care regimens including CarbPac (combination of Carboplatin, 75 mg/kg; Strem Chemicals, MA, and Paclitaxel, 15 mg/kg; Teva, CA); and GemCis (combination of Gemcitabine, 112 mg/kg; Eli Lilly, IN and Cisplatin, 4 mg/kg; Sigma, CA). Treatments were terminated after 8 weeks and lung lesions were analyzed.

m-CT (micro-computed tomography)

KrasG12D-LSL mice were administered with Adeno-Cre by inhalation as described. Mice received treatments including vehicle, PF-502, PD-901, and combination at 19.5 weeks post-inhalation and at doses described above. For m-CT imaging and at 4 weeks after treatment initiation, mice were anesthetized using isoflurane inhalation and were imaged at 15 min post-administration of ExiTron nano 12,000 (Miltenyi Biotech, CA). All images were acquired in standard resolution using respiratory gating (70 kVP, 114 μA, 300 ms integration time, 41 μm voxel size). Images were saved as 2D binary digital imaging communication (DICOM) files and later were analyzed using third-party DICOM image processing applications 64-bit OsiriX v.3.9.2 (Pixmeo; Switzerland) and/or Analyze v.10.0 (AnalyzeDirect; Overland Park, KS, USA).

Results

Combination of PF-502 and PD-901 inhibits tumor growth in H460 lung colonization model

The H460 cell line that carries mutations in Kras and PIK3CA has been widely used in a variety of in vitro and in vivo studies [26]. To test the efficacy of PF-502 and PD-901 in a NSCLC model and in order to monitor tumor progression and response to therapy, mice were iv injected with H460-Luc line. At day-7 post-injection, mice were randomized to treatment groups once Luc signal was detected and was quantified to select a cohort of mice for efficacy study (Fig. 1a). In addition to PF-502, PD-901, and combination of both agents, we also included two SOC (standard of care) regimens including CarbPac and GemCis in the treatment groups. Quantification of Luc signals in H460 tumors at three different time points post-treatment suggested that combination of PF-502 and PD-901 has greater tumor growth inhibition compared with either monotherapy (Fig. 1b). Our data also indicate that daily administration of PD-901 in the combination regimen at 1.5 mg/kg was not associated with any significant adverse effects or deaths during the course of treatment. Additionally, between the two SOC regimens, GemCis showed superior efficacy compared with CarbPac and the corresponding control (Fig. 1c). However, both SOC regimens were associated with some adverse effects in the recipients. In conclusion, our data in H460-Luc tumors are the first study to show that PD-901 is potent at sub-MTD dose to inhibit tumor growth and invasiveness when used in combination with an inhibitor of PI3K/mTOR.

Fig. 1
figure 1

a Scid/beige mice were iv implanted with human H460-Luc + cells (1 × 106 per mouse). Mice were imaged at day-7 post-implantation and were randomized to 7 groups (n = 10) receiving vehicle, PF-04691502, PD-0325901, combination of PF-04691502 and PD-0325901, injectable PBS, two standard of care regimens including CarbPac and GemCise. Treated animals were imaged in IVIS-100 system weekly for 3 consecutive weeks. b, c Quantification of photons of lungs in tumor bearing mice at last time point (day 27) treated with small molecule inhibitors (b) or standard of care (c)

Full size image

Combination of PF-502 and PD-901 at sub-MTD dosing significantly inhibited growth of adenocarcinoma lung tumors in KrasG12D-LSL mice

To further validate H460-Luc tumor data, we tested efficacy of combination regimen in KrasG12D-LSL mice using the same therapeutic regimens described for the H460-Luc tumor models. KrasG12D-LSL mice have been widely used for variety of tumor biology studies [8]. The conditional nature of KrasG12D activation allows tumor initiation in any tissue where Kras is a driver of tumor progression. Generation of lung tumors was one of the original studies in KrasG12D-LSL mice where intranasal inhalation of adenoviral expressing Cre recombinase (Adeno-Cre) resulted in the expression of KrasG12D in the lung and formation of adenomatous hyperplasia, papillary adenoma, large adenoma, and adenocarcinoma at different time points post-inhalation [8, 25]. Following the same methodology, we provided a strategy to develop tumors in KrasG12D mice and to treat them with our compounds. As depicted in Fig. 2a, KrasG12D mice were administered with Adeno-Cre (2.5 × 107) and lung lesions were analyzed in few mice at 8 weeks post-inhalation to ensure tumor initiation and progression. At 14 weeks post-inhalation when most of recipients have developed adenocarcinoma lesions [8, 25], mice were assigned to therapeutic regimens and the study was terminated at 22 weeks post-inhalation. Following the above strategy, and at 14 weeks post-inhalation, KrasG12D mice received the same treatments as H460-Luc study. To differentiate role of PI3K versus mTOR pathway, a selective mTOR inhibitor was also included in the regimens. Analysis of lung tissues showed that while PF-502 or PD-901 alone had the ability to inhibit growth of majority of adenocarcinoma lesions, combination therapy significantly suppressed adenocarcinoma lesions further suggesting that PD-901 at sub-MTD dose might have therapeutic benefit in NSCLC tumors (Fig. 2b). Similar to the results with the H460-Luc model, SOC regimens inhibited tumor growth in KrasG12D mice. Analysis of body weight in the treated animals did not reveal any significant difference between PD-901 (as single agent or in combination)- and vehicle-treated mice further suggesting lack of adverse effect when administering the compound at sub-MTD.

Fig. 2
figure 2

a A Strategy to investigate tumor progression and response to therapy in KrasG12D-LSL GEMMs. Mice would be administered (inhalation) with Adeno-Cre virus (2.5 × 107) as described and few mice would be analyzed at 8 weeks post-adeno-inhalation to ensure tumor progression in the lung. At 14 weeks post-inhalation, mice would be assigned to treatment with the compounds as described in the Methods. All the treated mice would be killed after 8 weeks (i.e. 22 weeks after Adeno-cre inhalation) and one lobe of lung tissue would be analyzed in histology using H&E staining. b Using our strategy, KrasG12D-LSL mice were administered with Adeno-Cre and received therapeutic compounds at 14 weeks post-inhalation followed by study termination and tissue analysis. Each image is representative of lung from one mouse, i.e., images of lung from 4 mice from each treatment are shown. Combination of PI3K/mTOR and MEK inhibitors has the most potent effect on tumor growth inhibition (reduced sub-gross lung tumor burden; manifested as dense, dark blue regions in the lung tissue) compared with all other treatments. c Body weights were measured in all the recipients once/week. Administration of PD-901 did not significantly change body weight versus all other treatments

Full size image

Combination of PF-502 and PD-901 significantly inhibits growth of all lung lesions

To further investigate detail of efficacy of combination regimen, lung lesions were classified as hyperplastic, benign adenoma, or adenocarcinoma and were quantified in each mouse and in each treatment in histopathology analyses (Fig. 3a). Interestingly, PSMI (selective mTOR inhibitor)-treated lungs had very few benign adenoma like lesions but there was no significant change in the total percentage of all lesions These data suggest that tumor initiation and progression in KrasG12D mice is independent of mTOR pathway. PF-502 as a single agent significantly inhibited the percentage of all three lesions in the lung compared with vehicle-treated animals. Strikingly, PD-901 as a single agent and at sub-MTD dose also significantly inhibited all three lesions in the lung further suggesting a strong potency for the compound even at levels significantly below MTD (Fig. 3b). Combination of PF-502 and PD-901 suppressed the percentage of all three lesions in the lung further validating human H460 tumor data. Lesion classification also allowed us to investigate frequency of each lesion in the recipients (Fig. 3c). Irrespective of type of treatment, all animals carried hyperplastic lesions indicating that none of the therapies can fully eradicate tumor initiation and progression in the recipients. Interestingly and despite efficacy as single agent in inhibiting percentage of adenoma lesions, PF-502 or PD-901 monotherapies did not completely inhibit the adenoma lesions. However, PSMI, combination regimen and SOCs reduced the percentage of mice carrying adenoma lesions. PD-901 as single agent or in combination with PF-502 showed superior efficacy in reducing frequency of mice carrying adenoma lesions. Finally, with the exception of vehicle and PSMI, all other treatments reduced percentage of mice carrying adenocarcinoma lesions. PD-901 as single agent or in combination with PF-502 showed superior efficacy in reducing frequency of mice with adenocarcinoma lesions (Fig. 3c). In conclusion, quantification of lesions further suggested a great potency for PD-901 at sub-MTD dose as single agent or in combination with PF-502.

Fig. 3
figure 3

Lesion classification to quantify percentage of lesions in lung in each mouse. KrasG12D-LSL mice were administered and treated as described in Fig. 2a. a Representative lesions (hyperplastic, benign adenoma, and adenocarcinoma) are shown in a lung tissue. b Percentage of lung occupied by each lesion type was quantified in each mouse and averaged for each treatment group. c Percentage of mice exhibiting each lesion in each treatment to investigate whether therapeutic compounds have the ability to fully inhibit tumor growth in the recipients. Bars represent percentage of mice bearing each lesion in the lung. While all the treated mice exhibit hyperplastic lesions, only 20 % of animals treated with combination of PI3K/mTOR and MEK inhibitors display adenocarcinoma lesions in the lung

Full size image

Combination of PF-502 and PD-901 significantly inhibits phosphorylation of ribosomal S6 (pS6) protein in KrasG12D-LSL mice

To investigate whether superior efficacy of PD-901 correlates with inhibition of downstream signaling pathway, phosphorylation status in ribosomal S6 protein was investigated in lung sections [19]. As illustrated in Fig. 4, pS6 signal was observed in majority of lesions in vehicle-treated tumors, indicative of pathway activation. PF-502 or PD-901 as single agent partially inhibited the signal in tumor cells; however, combination regimen fully inhibited pS6 in tumors further providing a correlation between tumor growth and pathway activation. These studies suggest that PD-901 potency and efficacy at sub-MTD dose is selectively associated with inhibition of the downstream signaling pathway.

Fig. 4
figure 4

KrasG12D-LSL mice were administered with Adeno-Cre and were treated as described above. Sections of lungs were stained with pS6 antibody using optimized protocol provided in the laboratory. Among all the treatments, combination of PF-502 and PD-901 has significantly inhibited S6 phosphorylation

Full size image

m-CT technology further verifies efficacy of combination of PF-502 and PD-901 therapy in live mice

One of the major caveats of adenovirus delivery of Cre recombinase is the lack of a reporter in the vector in order to horizontally monitor tumor progression in the mice. Additionally, given the transient nature of gene expression in any adenoviral expressing system, inclusion of any reporter in the adenoviral expressing Cre would not provide a permanent expression to monitor tumor growth at advance stages of tumorigenesis. For these reasons, development of other imaging modalities would provide a great approach in monitoring tumor progression and response to the therapies. Among several imaging techniques, m-CT has proven to be a reliable method to image lung tumors at different stages of tumorigenesis [27]. To further confirm efficacy of PD-901 at sub-MTD dose as single agent or in combination regimen and to further gain confidence in efficacy study, we treated mice at ~19.5 weeks after inhalation where majority of lung is covered by tumor lesions. Image analysis of whole lungs at 4 weeks after the treatment clearly shows tumor regions in the vehicle-treated mice compared with non-inhaled animals (Fig. 5a). Consistent with histology data, PF-502 partially inhibited tumor growth, whereas PD-901 showed a better efficacy as monotherapy (Fig. 5b). Combination regimen provided a superior efficacy compared with monotherapies and further confirmed histology and lesion classification data (Fig. 5b, c).

Fig. 5
figure 5

m-CT images further support efficacy of PD-901 as single agent or in combination regimen in treated animals. KrasG12D-LSL mice were administered with Adeno-Cre as described and received treatments (vehicle, PF-502, PD-901 and PF-502 + PD901) at ~ 19.5 weeks post-inhalation. M-CT images were taken from control wild type (a) and KrasG12D-LSL mice (b) at 4 weeks after the treatment. c Quantification of tumor volume (relative to vehicle) in control wild type and treated mice is included

Full size image

Discussion

Here, we examined the efficacy of PD-901 at a sub-MTD dose in two different lung tumor models where conditional/spontaneous activation of mutated Kras (in both KrasG12D-LSL mice and H460 tumor cells) and PIK3CA (in H460 tumor cells) are major drivers of tumorigenesis. In order to mimic clinical status where most patients are diagnosed and receive therapies at advance stage of tumorigenesis, treatments started when tumors were established in H460 model or at advanced stage (adenocarcinoma) in KrasG12D-LSL GEM. Efficacy data as well as animals overall health (as measured by body weight) suggested an excellent potency for PD-901 as a single agent or in combination with PF-502 that was well-tolerated by the recipients. Combination therapy significantly suppressed tumor growth in both H460-Luc and KrasG12D-LSL tumors; however, PD-901, as a single agent, had superior efficacy in the latter suggesting that PIK3CA is a major driver of tumorigenesis in the former. Comparison of two SOCs also showed that the overall efficacy of CarbPac was greater in both models compared with GemCis further raising a question about the mechanism of action in lung tumors between the two regimens. Detailed analysis of lesions in the KrasG12D-LSL tumors showed that inhibition of mTOR alone may not provide a benefit, while the dual inhibitor (PF-502) significantly inhibited growth of all 3 lesions in the lung. Future studies using a selective PI3K inhibitor will determine whether, compared with dual inhibitors, suppression of PI3K is sufficient to inhibit tumor growth in KrasG12D-LSL mice. Interestingly, PD-901 as a single agent was very efficacious in inhibiting tumor progression in KrasG12D-LSL mice further emphasizing the significance of targeting the driver of tumorigenesis (i.e., MEK pathway) in this model. Analysis of lung lesions in individual mice showed that none of the therapies eliminated the development of hyperplastic lesions further confirming that tumors may not be fully prevented or reversed by any treatment. Adenoma and adenocarcinoma lesions were most affected by combination regimen indicating an active cross-talk between MEK and PI3K at this stage of tumorigenesis. To investigate whether downstream signaling pathway is also targeted by the therapies, IHC data showed that pS6 is almost fully inhibited only in the combination group correlating efficacy data and activity of signaling pathways in KrasG12D-LSL tumors. Finally, m-CT imaging of live mice treated with PD-901 and PF-502 at an advance stage of tumor progression (~19.5 weeks post-inhalation) exhibited reduced lesion size, further supporting IHC data. Overall, our investigations indicate that PD-901 is a potent compound at doses significantly below the established MTD but without previously reported major adverse effects in the recipients and suggest that therapeutic compounds might be very efficacious at lower dose if they are administered in cancers with targeted mutations/pathway activation. Future studies will determine clinical benefit of combination of PF-502 and PD-901 in cancer patients.