ErbB2-positive mammary tumors can escape PI3K-p110α loss through downregulation of the Pten tumor suppressor

Abstract

Breast cancer is the most common cancer among women and 30% of patients will be diagnosed with an ErbB2-positive tumor. Forty percent of ErbB2-positive breast tumors have an activating mutation in p110α, a catalytic subunit of phosphoinositide 3-kinase. Clinical and experimental data show that breast tumors treated with a p110α-specific inhibitor often circumvent inhibition and resume growth. To understand this mechanism of resistance, we crossed a p110α conditional (p110α^flx/flx) mouse model with mice that overexpress the ErbB2/Neu-IRES-Cre transgene (NIC) specifically in the mammary epithelium. Although mammary-specific deletion of p110α dramatically delays tumor onset, tumors eventually arise and are dependent on p110β. Through biochemical analyses we find that a proportion of p110α-deficient tumors (23%) display downregulation of the Pten tumor suppressor. We further demonstrate that loss of one allele of PTEN is sufficient to shift isoform dependency from p110α to p110β in vivo. These results provide insight into the molecular mechanism by which ErbB2-positive breast cancer escapes p110α inhibition.

Figure 1. Loss of p110α significantly delays tumor onset, and impairs tumor growth and metastasis, with a subset of tumors exhibiting down regulation of the Pten tumor suppressor

(a) Kaplan-Meier tumor onset curve for NIC mice that are wildtype, heterozygous, or homozygous for the p110α conditional allele. The table indicates for each genotype the penetrance (percentage of animals that developed tumours), T₅₀ (age when 50% of the animals have tumours), and average tumour onset with standard deviation for each of the curves shown on the graph. p values were calculated using a two-tailed student t-test. (b) The number of tumors and total tumor burden at endpoint (5-7 weeks post-palpation). The error bars represent the standard error of the mean and the p values were calculated using a two-tailed student t-test. (c) immunoblot analysis of tumor lysates (20μg) for the genotypes indicated.

Figure 2. p110α-deficient ErbB2 tumors remain dependent on PI3K

(a) Immunoblot of p110α^wt/wt/NIC mammary tumor lysates (20μg) from NCr mice injected with (500 000 tumor cells) after treatment with either a vehicle (veh) 125mg/kg of GDC-0941 by oral gavage for the indicated times, (b) Tumor outgrowth in NCr mice injected with tumor cells (500 000 tumor cells) of the indicated genotypes and treated with either vehicle or 125mg/kgof GDC-0941 by oral gavage daily for 6 weeks (c) Difference in tumor growth in between vehicle-treated and GDC-0941-treated tumors after 3 weeks of treatment. The p-values were calculated using a two-tailed Student t-test. (d) Immunoblot of NCr tumor lysates (20μg) from the indicated genotypes.

Figure 3. PTEN haploinsufficiency can compensate for loss of p110α in ErbB2 mammary tumor progression

(a) Kaplan-Meier tumor onset curve for NIC mice that are wild-type or homozygous for the p110α conditional allele and wild type or heterozygous for the PTEN conditional allele. The table indicates for each genotype the penetrance (percentage of animals that developed tumors), T₅₀ (age when 50% of the animals have tumors), and average tumor onset with standard deviation for each of the curves shown on the graph, p-values were calculated using the log-rank (Mantel-Cox) test, (b) The number of tumors and total tumor burden at endpoint (5-7 weeks post-palpation). The error bars represent the standard error of the mean and the p-values were calculated using a two-tailed student t-test (c) Immunoblot analysis of tumor lysates (20μg) for the genotypes indicated.

Figure 4. Tumors that have lost p110α are p110β dependent

(a) Immunoblot of NCr cells treated with vehicle or 100nM, 500nM or 1000nM of TGX-221 (p110β3 specific inhibitor) for 6 hours, (b) Co-immunoprecipitation of p110βand ErbB3, p110βand p-85 (500μg). (c) Immunoflorescence staining of p110βand DAPI on paraffin-embedded mammary tumor tissue, scale bar represents 50μm.

Figure 5. Summary diagram of P13K signaling in response to p110α targeted therapy

(a) PI3K signaling in ErbB2 positive breast cancer before therapy, where the majority of PDK signaling occurs through p110α binding to the ErbB3/ErbB2 heterodimer. (b) PI3Ksignaling after treatment with a p110α-specific inhibitor. In response to p110α a subset of tumors escape through p110β in response to PTEN downregulation/heterozygous loss. In these tumors p110β may signal through theErbB2/ErbB3 heterodimer but is more likely to be signaling through another receptor, possibly GPCR's or lntegrins.