Integrated CRISPR screening and drug profiling identifies combination opportunities for EGFR, ALK, and BRAF/MEK inhibitors

Anti-tumor efficacy of targeted therapies is variable across patients and cancer types. Even in patients with initial deep response, tumors are typically not eradicated and eventually relapse. To address these challenges, we present a systematic screen for targets that limit the anti-tumor efficacy of EGFR and ALK inhibitors in non-small cell lung cancer and BRAF/MEK inhibitors in colorectal cancer. Our approach includes genome-wide CRISPR screens with or without drugs targeting the oncogenic driver ("anchor therapy"), and large-scale pairwise combination screens of anchor therapies with 351 other drugs. Interestingly, targeting of a small number of genes, including MCL1, BCL2L1, and YAP1, sensitizes multiple cell lines to the respective anchor therapy. Data from drug combination screens with EGF816 and ceritinib indicate that dasatinib and agents disrupting microtubules act synergistically across many cell lines. Finally, we show that a higher-order-combination screen with 26 selected drugs in two resistant EGFR-mutant lung cancer cell lines identified active triplet combinations.

A conditional inducible JAK2V617F transgenic mouse model reveals myeloproliferative disease that is reversible upon switching off transgene expression

Aberrant activation of the JAK/STAT pathway is thought to be the critical event in the pathogenesis of the chronic myeloproliferative neoplasms, polycythemia vera, essential thrombocythemia and primary myelofibrosis. The most frequent genetic alteration in these pathologies is the activating JAK2^V617F mutation, and expression of the mutant gene in mouse models was shown to cause a phenotype resembling the human diseases. Given the body of genetic evidence, it has come as a sobering finding that JAK inhibitor therapy only modestly suppresses the JAK2^V617F allele burden, despite showing clear benefits in terms of reducing splenomegaly and constitutional symptoms in patients. To gain a better understanding if JAK2^V617F is required for maintenance of myeloproliferative disease once it has evolved, we generated a conditional inducible transgenic JAK2^V617F mouse model using the SCL-tTA-2S tet-off system. Our model corroborates that expression of JAK2^V617F in hematopoietic stem and progenitor cells recapitulates key hallmarks of human myeloproliferative neoplasms, and exhibits gender differences in disease manifestation. The disease was found to be transplantable, and importantly, reversible when transgenic JAK2^V617F expression was switched off. Our results indicate that mutant JAK2^V617F-specific inhibitors should result in profound disease modification by disabling the myeloproliferative clone bearing mutant JAK2.

Identifying optimal biologic doses of everolimus (RAD001) in patients with cancer based on the modeling of preclinical and clinical pharmacokinetic and pharmacodynamic data

PURPOSE

To use preclinical and clinical pharmacokinetic (PK)/pharmacodynamic (PD) modeling to predict optimal clinical regimens of everolimus, a novel oral mammalian target of rapamycin (mTOR) inhibitor, to carry forward to expanded phase I with tumor biopsy studies in cancer patients.

PATIENTS AND METHODS

Inhibition of S6 kinase 1 (S6K1), a molecular marker of mTOR signaling, was selected for PD analysis in peripheral blood mononuclear cells (PBMCs) in a phase I clinical trial. PK and PD were measured up to 11 days after the fourth weekly dose. A PK/PD model was used to describe the relationship between everolimus concentrations and S6K1 inhibition in PBMCs of cancer patients and in PBMCs and tumors of everolimus-treated CA20948 pancreatic tumor-bearing rats.

RESULTS

Time- and dose-dependent S6K1 inhibition was demonstrated in human PBMCs. In the rat model, a relationship was shown between S6K1 inhibition in tumors or PBMCs and antitumor effect. This allowed development of a direct-link PK/PD model that predicted PBMC S6K1 inhibition-time profiles in patients. Comparison of rat and human profiles simulated by the model suggested that a weekly 20- to 30-mg dose of everolimus would be associated with an antitumor effect in an everolimus-sensitive tumor and that daily administration would exert a greater effect than weekly administration at higher doses.

CONCLUSION

A direct-link PK/PD model predicting the time course of S6K1 inhibition during weekly and daily everolimus administration allowed extrapolation from preclinical studies and first clinical results to select optimal doses and regimens of everolimus to explore in future clinical trials.

Dual inhibition of mTOR and estrogen receptor signaling in vitro induces cell death in models of breast cancer

PURPOSE

RAD001 (everolimus), a mammalian target of rapamycin (mTOR) pathway inhibitor in phase II clinical trials in oncology, exerts potent antiproliferative/antitumor activities. Many breast cancers are dependent for proliferation on estrogens synthesized from androgens (i.e., androstenedione) by aromatase. Letrozole (Femara) is an aromatase inhibitor used for treatment of postmenopausal women with hormone-dependent breast cancers. The role of the mTOR pathway in estrogen-driven proliferation and effects of combining RAD001 and letrozole were examined in vitro in two breast cancer models.

EXPERIMENTAL DESIGN

The role of the mTOR pathway in estrogen response was evaluated in aromatase-expressing MCF7/Aro breast cancer cells by immunoblotting. Effects of RAD001 and letrozole (alone and in combination) on the proliferation and survival of MCF7/Aro and T47D/Aro cells were evaluated using proliferation assays, flow cytometry, immunoblotting, and apoptosis analyses.

RESULTS

Treatment of MCF7/Aro cells with estradiol or androstenedione caused modulation of the mTOR pathway, a phenomenon reversed by letrozole or RAD001. In MCF7/Aro and T47D/Aro cells, both agents inhibited androstenedione-induced proliferation; however, in combination, this was significantly augmented (P < 0.001, two-way ANOVA, synergy by isobologram analysis). Increased activity of the combination correlated with more profound effects on G1 progression and a significant decrease in cell viability (P < 0.01, two-way ANOVA) defined as apoptosis (P < 0.05, Friedman test). Increased cell death was particularly evident with optimal drug concentrations.

CONCLUSION

mTOR signaling is required for estrogen-induced breast tumor cell proliferation. Moreover, RAD001-letrozole combinations can act in a synergistic manner to inhibit proliferation and trigger apoptotic cell death. This combination holds promise for the treatment of hormone-dependent breast cancers.

Everolimus suppresses cancellous bone loss, bone resorption, and cathepsin K expression by osteoclasts

The proliferation inhibitor of the macrolide class, everolimus, is a drug shown to be effective in the prevention of organ transplant rejection and to have a potential in the treatment of rheumatoid arthritis and certain cancers. As these diseases or their current treatments are associated with bone loss, we examined the effect of everolimus on mouse and human bone cells in vitro and on bone in an ovariectomized (OVX) rat model. Everolimus potently inhibited primary mouse and human osteoclast activity in the pit assay (IC50 values of 0.6-4.0 nM), as well as osteoclast formation, measured as the number of tartrate-resistant acid phosphatase (TRAP) multinucleated cells (IC50 values of 7.7-10.5 nM). Inhibition of osteoblastic differentiation was also observed (IC50 value of 13.5 nM). As expected, everolimus inhibited proliferation of osteoclast precursors and stimulated apoptosis, albeit with insufficient potency and efficacy to explain inhibition of osteoclast activity. Thus, everolimus appeared to directly inhibit bone resorption, which is in accord with the detected inhibition of mRNA and protein expression of cathepsin K; the main collagen-degrading protease in osteoclasts. Despite the in vitro antiproliferative activity of everolimus and the observed inhibition of osteoblast differentiation, no detrimental effects were detected at different skeletal sites in mature OVX rats at doses up to 3 mg/kg/day. This everolimus dose also prevented the OVX-induced loss of cancellous bone by 60%, an effect predominantly associated with decreased osteoclast-mediated bone resorption, resulting in a partial preservation of the cancellous bone network. Everolimus inhibited S6 kinase 1 activity in rat blood cells, skin, and bone, at doses equivalent to those used for efficacy experiments in the OVX rat model, which demonstrated in vivo targeting of the expected molecular pathway. In conclusion, everolimus directly inhibits bone resorption by osteoclasts and thus could at least be neutral or protective for bone in vivo, which would favor its use in disease indications associated with bone loss.

Antitumor efficacy of intermittent treatment schedules with the rapamycin derivative RAD001 correlates with prolonged inactivation of ribosomal protein S6 kinase 1 in peripheral blood mononuclear cells

The orally bioavailable rapamycin derivative RAD001 (everolimus) targets the mammalian target of rapamycin pathway and possesses potent immunosuppressive and anticancer activities. Here, the antitumor activity of RAD001 was evaluated in the CA20948 syngeneic rat pancreatic tumor model. RAD001 demonstrated dose-dependent antitumor activity with daily and weekly administration schedules; statistically significant antitumor effects were observed with 2.5 and 0.5 mg/kg RAD001 administered daily [treated tumor versus control tumor size (T/C), 23% and 23-30%, respectively], with 3-5 mg/kg RAD001 administered once weekly (T/C, 14-36%), or with 5 mg/kg RAD001 administered twice weekly (T/C, 36%). These schedules were well tolerated and exhibited antitumor potency similar to that of the cytotoxic agent 5-fluorouracil (T/C, 23%). Moreover, the efficacy of intermittent treatment schedules suggests a therapeutic window allowing differentiation of antitumor activity from the immunosuppressive properties of this agent. Detailed biochemical profiling of mammalian target of rapamycin signaling in tumors, skin, and peripheral blood mononuclear cells (PBMCs), after a single administration of 5 mg/kg RAD001, indicated that RAD001 treatment blocked phosphorylation of the translational repressor eukaryotic initiation factor 4E-binding protein 1 and inactivated the translational activator ribosomal protein S6 kinase 1 (S6K1). The efficacy of intermittent treatment schedules was associated with prolonged inactivation of S6K1 in tumors and surrogate tissues (> or =72 h). Furthermore, detailed analysis of the dose dependency of weekly treatment schedules demonstrated a correlation between antitumor efficacy and prolonged effects (> or =7 days) on PBMC-derived S6K1 activity. Analysis of human PBMCs revealed that S6K1 also underwent a concentration-dependent inactivation after RAD001 treatment ex vivo (>95% inactivation with 20 nM RAD001). In contrast, human PBMC-derived eukaryotic initiation factor 4E-binding protein 1 was present predominantly in the hypophosphorylated form and was unaffected by RAD001 treatment. Taken together, these results demonstrate a correlation between the antitumor efficacy of intermittent RAD001 treatment schedules and prolonged S6K1 inactivation in PBMCs and suggest that long-term monitoring of PBMC-derived S6K1 activity levels could be used for assessing RAD001 treatment schedules in cancer patients.

Selective in vivo and in vitro effects of a small molecule inhibitor of cyclin-dependent kinase 4

BACKGROUND

Cyclin-dependent kinase 4 (Cdk4) represents a prime target for the treatment of cancer because most human cancers are characterized by overexpression of its activating partner cyclin D1, loss of the natural Cdk4-specific inhibitor p16, or mutation(s) in Cdk4's catalytic subunit. All of these can cause deregulated cell growth, resulting in tumor formation. We sought to identify a small molecule that could inhibit the kinase activity of Cdk4 in vitro and to then ascertain the effects of that inhibitor on cell growth and tumor volume in vivo.

METHODS

A triaminopyrimidine derivative, CINK4 (a chemical inhibitor of Cdk4), was identified by screening for compounds that could inhibit Cdk4 enzyme activity in vitro. Kinase assays were performed on diverse human Cdks and on other kinases that were expressed in and purified from insect cells to determine the specificity of CINK4. Cell cycle effects of CINK4 on tumor and normal cells were studied by flow cytometry, and changes in phosphorylation of the retinoblastoma protein (pRb), a substrate of Cdk4, were determined by western blotting. The effect of the inhibitor on tumor growth in vivo was studied by use of tumors established through xenografts of HCT116 colon carcinoma cells in mice. Statistical tests were two-sided.

RESULTS

CINK4 specifically inhibited Cdk4/cyclin D1 in vitro. It caused growth arrest in tumor cells and in normal cells and prevented pRb phosphorylation. CINK4 treatment resulted in statistically significantly (P: =.031) smaller mean tumor volumes in a mouse xenograft model.

CONCLUSIONS

Like p16, the natural inhibitor of Cdk4, CINK4 inhibits Cdk4 activity in vitro and slows tumor growth in vivo. The specificity of CINK4 for Cdk4 raises the possibility that this small molecule or one with a similar structure could have therapeutic value.

Expression and purification of human recombinant GST-FGF receptor-1

When tumors undergo the angiogenic switch, cell growth and tissue invasion is facilitated by the formation of new capillaries from preexisting blood vessels, a process known as angiogenesis. Growth factors such as vascular endothelial growth factor/vascular permeability factor (VEGF/VPF) and fibroblast growth factor (FGF) trigger the process of angiogenesis. Here we describe a protocol for the expression and one-step purification of human recombinant GST-FGF receptor type 1 (FGFR-1) from Sf9 cells. This protocol allows generating an active kinase as indicated by its reactivity with a monoclonal antibody to phosphorylated tyrosine. The purified enzyme displays a specific activity of 1.2 x 10(4) pmol mg(-1) min(-1), which is in the range of activities reported for homogeneously purified recombinant kinases. We have employed a number of compounds to show that the GST-FGFR-1 preparation is suitable to the identification of tyrosine kinase inhibitors. Considering that inhibitors of angiogenesis may represent an attractive tool in therapeutic strategies targeting invasive metastatic tumors the results presented here, along with available data on the structure of the ATP-binding pocket of FGFR-1, should facilitate the rational design of specific FGFR-1 inhibitory compounds.

Inhibition of cyclin-dependent kinase 4 (Cdk4) by fascaplysin, a marine natural product

Small chemical molecules that interfere with biological proteins could be useful for gaining insight into the complex biochemical processes in mammalian cells. Cdk4 is a key protein whose activity is required not only for emergence of cells from quiescence but also at the G1/S transition in the cell cycle and which is misregulated in 60-70% of human cancers. We set out to identify chemical inhibitors of Cdk4 and discovered that, in vitro, fascaplysin specifically inhibited Cdk4. Molecular modelling based on the crystal structure of Cdk2 suggests that fascaplysin inhibits Cdk4 by binding to the ATP pocket of the kinase. Treatment of tumour (p16(-), pRb(+)) and normal (p16(+), pRb(+)) cell lines with fascaplysin caused G1 arrest and prevented pRb phosphorylation at sites implicated as being specific for Cdk4 kinase. Fascaplysin will therefore prove to be a useful tool in studying the consequence of Cdk4 inhibition, especially in cells containing inactivated p16.

DNA binding protein dbpA binds Cdk5 and inhibits its activity

Progress in the cell cycle is governed by the activity of cyclin dependent kinases (Cdks). Unlike other Cdks, the Cdk5 catalytic subunit is found mostly in differentiated neurons. Interestingly, the only known protein that activates Cdk5 (i.e. p35) is expressed solely in the brain. It has been suggested that, besides its requirement in neuronal differentiation, Cdk5 activity is induced during myogenesis. However, it is not clear how this activity is regulated in the pathway that leads proliferative cells to differentiation. In order to find if there exists any Cdk5-interacting protein, the yeast two-hybrid system was used to screen a HeLa cDNA library. We have determined that a C-terminal 172 amino acid domain of the DNA binding protein, dbpA, binds to Cdk5. Biochemical analyses reveal that this fragment (dbpA(Cdelta)) strongly inhibits p35-activated Cdk5 kinase. The protein also interacts with Cdk4 and inhibits the Cdk4/cyclin D1 enzyme. Surprisingly, dbpA(Cdelta) does not bind Cdk2 in the two-hybrid assay nor does it inhibit Cdk2 activated by cyclin A. It could be that dbpA's ability to inhibit Cdk5 and Cdk4 reflects an apparent cross-talk between distinct signal transduction pathways controlled by dbpA on the one hand and Cdk5 or Cdk4 on the other.

Identification of a human cDNA encoding a kinase-defective cdk5 isoform

The cyclin-dependent kinase 5 (Cdk5) catalytic subunit is expressed in both cycling and noncycling cells and is present in many tissues. Neuronal and muscle cells contain the highest amount of this protein. The p35 protein, which is expressed solely in the brain, activates Cdk5. Cdk5 activity is involved in terminal differentiation of neurons and muscle cells. We attempted to clone cdk5 by PCR from a human fetal brain cDNA library. Surprisingly, we amplified two forms of the cdk5 gene, the wild type and a cdk5 variant that lacks the complete kinase domain VI. The variant is also found in SH-SY-5Y neuroblastoma cells but not in T-cells, HeLa cells, the thymus, and placental tissue. The protein encoded by the cdk5 variant, the Cdk5 isoform (Cdk5i), purifies with p35 when coexpressed in insect cells. The activity associated with the heterodimer Cdk5i/p35 is found to be appreciably weaker than the wild-type Cdk5/p35 kinase. Moreover, Cdk5i/p35 cannot autophosphorylate its two subunits as with Cdk5/p35. Interestingly, kinase-defective Cdk5i can abolish the activity of wild-type Cdk5 when both are coexpressed with p35 in insect cells, suggesting that Cdk5i may have a function in regulating Cdk5 activity in human cells too.