Preclinical evaluation of novel fatty acid synthase inhibitors in primary colorectal cancer cells and a patient-derived xenograft model of colorectal cancer

Abstract

Fatty Acid Synthase (FASN), a key enzyme of de novo lipogenesis, is upregulated in many cancers including colorectal cancer (CRC); increased FASN expression is associated with poor prognosis. Potent FASN inhibitors (TVBs) developed by 3-V Biosciences demonstrate anti-tumor activity in vitro and in vivo and a favorable tolerability profile in a Phase I clinical trial. However, CRC characteristics associated with responsiveness to FASN inhibition are not fully understood. We evaluated the effect of TVB-3664 on tumor growth in nine CRC patient-derived xenografts (PDXs) and investigated molecular and metabolic changes associated with CRC responsiveness to FASN inhibition. CRC cells and PDXs showed a wide range of sensitivity to FASN inhibition. TVB-3664 treatment showed significant response (reduced tumor volume) in 30% of cases. Anti-tumor effect of TVB-3664 was associated with a significant decrease in a pool of adenine nucleotides and alterations in lipid composition including a significant reduction in fatty acids and phospholipids and an increase in lactosylceramide and sphingomyelin in PDXs sensitive to FASN inhibition. Moreover, Akt, Erk1/2 and AMPK were major oncogenic pathways altered by TVBs. In summary, we demonstrated that novel TVB inhibitors show anti-tumor activity in CRC and this activity is associated with a decrease in activation of Akt and Erk1/2 oncogenic pathways and significant alteration of lipid composition of tumors. Further understanding of genetic and metabolic characteristics of tumors susceptible to FASN inhibition may enable patient selection and personalized medicine approaches in CRC.

Keywords: FASN; TVB-3664; colorectal cancer; lipogenesis; patient-derived xenografts.

Figure 1. CRC cell lines exhibit a wide range of sensitivity to FASN inhibition

(A) A panel of CRC cell lines was treated with TVB-3166 (0.2 μM) for 7 days in their normal medium supplemented with 10% FBS. Medium and drug were changed on day 4. Proliferation was determined by cell count. Data shown as percent response to FASN inhibition. (B) The two most sensitive (CaCo2 and HT29) and the most resistant (LIM2405) cell lines were treated with multiple concentrations of TVB-3664 for 7 days without medium change and the number of cells was counted (^*p < 0.05). (C, E) Primary CRC cells Pt 93 and Pt 130 were treated with multiple concentrations of TVBs for 7 days and proliferation was determined by cell count. ^*p < 0.05. (D, F) Primary CRC cells were treated for 24 h (Pt 93) and 7 days (Pt 93 and Pt 130) with TVB inhibitors and cell lyses were subjected to western blot analysis.

Figure 2. Effect of TVB-3664 on tumor growth in CRC PDX models

(A) Schematic representation of study design. (B) Representative IHC images taken from resected patient tumor tissues stained for FASN expression. (C) PDX models sensitive to TVB-3664 treatment. Tumor response to FASN inhibition is shown as a fold change in tumor volume over time. Middle row shows corresponding tumor size at the end of the experiment. Bottom row shows corresponding fold change in weight of animals over time. Animals were treated daily with 3 mg/kg (Pt 2614 and Pt 2449PT) or 6 mg/kg (Pt 2402 and Pt 2449LM) of TVB-3664 by oral gavage (^*p < 0.05; Pt 2377PT and Pt 2377LM are established from matched primary and liver metastasis tissues from Pt 2377). No response to TVB-3664 treatment was observed in PDX Pt 2449LM. (D) PDX models resistant to FASN inhibition. Tumor response to FASN inhibition is shown as a fold change in tumor volume over time. Middle row shows corresponding tumor size at the end of the experiment. Bottom row shows corresponding fold change in weight of animals over time. Animals were treated daily with 3 mg/kg of TVB-3664 by oral gavage (^*p < 0.05). (E) Accelerated tumor growth in Pt 2377 PDX models treated with 3 mg/kg of TVB-3664. Tumor response to FASN inhibition is shown as a fold change in tumor volume over time in PDX models established from primary CRC (left) and liver metastasis tumors (right) Pt 2377. Middle row shows corresponding tumor size at the end of the experiment. Bottom row shows corresponding fold change in weight of animals over time (^*p < 0.05).

Figure 3. Expression of FASN and activation of FASN-associated oncogenic pathways in PDX models

(A) Western blot analysis of tissues from PDX models (all G1; Pt 2387 G2) showing expression/activation of major oncogenic pathways associated with de novo lipogenesis. Cases sensitive to FASN inhibition are shown in bold (arrows). (B) Western blot analysis of tumor tissues from Pt 2402 treated with 6 mg/kg of TVB-3664 daily for 5 weeks. (C) Western blot analysis of tissues the Pt 2387 model treated with 3 mg/kg of TVB-3664 daily for 5 weeks. (D) Western blot analysis of tumor tissues from Pt 2377PT and LM treated with 3 mg/kg of TVB-3664 daily for 4 weeks.

Figure 4. Inhibition of FASN alters metabolites levels in PDX models sensitive to TVB-3664

(A) Changes in plasma metabolites in Pt 2402, Pt 2449PT and Pt 2614 PDXs. (B) Representative figures of NMR spectra obtained from analysis of plasma from the Pt 2402 model (control vs TVB-3664 treated). Levels of AXP (C) and m-Ins-2 (D) in tumor tissues from PDX models sensitive to TVB-3664 (p < 0.05) (see Methods).

Figure 5. Inhibition of FASN alters lipid composition in PDX models sensitive to TVB-3664

(A) Total plasma palmitate levels in control and TVB-3664-treated mice measured by mass spectrometry. FA (50 μl of a 1 μM C17) was added to 50 μl of plasma as an internal standard. Half of each sample was saponificated and derivatized and total palmitate measured by mass spectrometry (^*p < 0.05). (B–C) FASN-mediated changes in lipid classes common among Pt 2402, Pt 2449PT and Pt 2614 PDX models. Lipid classes were evaluated by grouping the lipids to a class, and within each class setting lipids with fold-change >0 as successes, and <0 as failures, and testing the ratio of successes to failures to 0.5 using a two-sided binomial test. The reported value for the binomial test is the log-ratio of the calculated proportion of successes over 0.5 (see Methods).