3'-Deoxy-3'-18F-fluorothymidine PET predicts response to (V600E)BRAF-targeted therapy in preclinical models of colorectal cancer

Abstract

Selective inhibition of oncogenic targets and associated signaling pathways forms the basis of personalized cancer medicine. The clinical success of (V600E)BRAF inhibition in melanoma, coupled with the emergence of acquired resistance, underscores the importance of rigorously validating quantitative biomarkers of treatment response in this and similar settings. Because constitutive activation of BRAF leads to proliferation in tumors, we explored 3'-deoxy-3'-(18)F-fluorothymidine ((18)F-FLT) PET to noninvasively quantify changes in tumor proliferation that are associated with pharmacologic inhibition of (V600E)BRAF downstream effectors and that precede changes in tumor volume.

Methods: Human colorectal cancer (CRC) cell lines expressing (V600E)BRAF were used to explore relationships between upregulation of p27 and phosphorylation of BRAF downstream effectors on small-molecule (V600E)BRAF inhibitor exposure. Athymic nude mice bearing (V600E)BRAF-expressing human CRC cell line xenografts were treated with a small-molecule (V600E)BRAF inhibitor (or vehicle) daily for 10 d. Predictive (18)F-FLT PET was conducted before changes in tumor volume occurred. Correlations were evaluated among PET, inhibition of phosphorylated MEK (p-MEK) and phosphorylated-ERK (p-ERK) by Western blot, tumor proliferation by histology, and small-molecule exposure by matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry (IMS).

Results: Treatment of CRC cell lines with PLX4720 reduced proliferation associated with target inhibition and upregulation of p27. In vivo, PLX4720 treatment reduced (18)F-FLT uptake, but not (18)F-FDG uptake, in Lim2405 xenografts before quantifiable differences in xenograft volume. Reduced (18)F-FLT PET reflected a modest, yet significant, reduction of Ki67 immunoreactivity, inhibition of p-MEK and p-ERK, and elevated tumor cell p27 protein levels. Both (18)F-FLT PET and (18)F-FDG PET accurately reflected a lack of response in HT-29 xenografts, which MALDI imaging mass spectrometry suggested may have stemmed from limited PLX4720 exposure.

Conclusion: We used preclinical models of CRC to demonstrate (18)F-FLT PET as a sensitive predictor of response to (V600E)BRAF inhibitors. Because (18)F-FLT PET predicted reduced proliferation associated with attenuation of BRAF downstream effectors, yet (18)F-FDG PET did not, these data suggest that (18)F-FLT PET may represent an alternative to (18)F-FDG PET for quantifying clinical responses to BRAF inhibitors.

Figure 1. In vitro PLX4720 exposure decreases proliferation and leads to elevated p27 levels in human CRC cell lines harboring ^V600EBRAF

Lim2405 and HT-29 cells expressing ^V600EBRAF exhibited marked sensitivity to PLX4720. (A) PI flow cytometry demonstrated concentration-dependent G1-arrest and decreased S-phase fractions. Western blot analysis of lysates collected from HT-29 (B) and Lim2405 (C) cells following 48 hr in vitro PLX4720 exposure (0, 10 nM, 100 nM, 250 nM, 1 μM, 5 μM) are shown. In both cell lines, p-MEK was inhibited at concentrations greater than 250 nM. Additionally, increased p27 protein expression was observed at PLX4720 concentrations greater than 250 nM.

Figure 2. Reduced in vivo ¹⁸F-FLT PET precedes quantifiable, PLX4720-dependent differences in tumor volume and correlates with drug exposure in Lim2405 xenografts

Mice bearing Lim2405 xenografts were treated daily with PLX4720 or vehicle. (A) Over a 10 day regimen, PLX4720 treatment induced significant xenograft growth arrest detectable by treatment day 7. ¹⁸F-FLT PET was carried out on treatment day 4 when there was negligible difference in tumor volume between PLX4720-treated and vehicle-treated controls (blue box). (B) Representative transverse images of vehicle-treated and PLX4720-treated cohorts collected on treatment day 4 (xenograft denoted by arrowhead). (C) ¹⁸F-FLT PET was significantly lower in PLX4720-treated xenografts compared to vehicle-treated controls. Immediately following PET, mice were sacrificed and thin, transverse sections of vehicle-treated and PLX4720-treated mice were prepared for MALDI IMS. (D) Pseudocolor maps of ion intensity corresponding to the parent molecular mass of PLX4720 demonstrate minimal signal intensity in vehicle-treated animals. In contrast, treated animals demonstrated considerable PLX4720 ion intensity, with the drug primarily localized to tumor tissue (dotted line). Other tissues exhibiting PLX4720 ion intensity included the bowel, consistent with oral administration.

Figure 3. Early ¹⁸F-FLT PET predicts inhibition of BRAF downstream effectors and p27 up-regulation

Western blot analysis quantifying BRAF effector molecules and p27 in ¹⁸F-FLT PET imaging-matched Lim2405 xenografts (see Figure 2). Compared to vehicle controls, xenografts harvested from PLX4720-treated cohorts at treatment day 4 exhibited reduced levels of p-MEK and p-ERK and elevated levels of p27.

Figure 4. Early ¹⁸F-FLT PET agrees with tumor cell TK1 levels and Ki67 immunohistochemistry

Immunohistochemistry analysis of ¹⁸F-FLT PET imaging-matched Lim2045 xenografts (see Figure 2). (A) Representative TK1 and Ki67 immunohistochemistry images of xenografts harvested at treatment day 4 are shown for both vehicle-treated and PLX720-treated tumors. (B) PLX4720-treated xenografts exhibited approximately 25% fewer TK1-positive cells per field compared to vehicle controls (34.3% vs. 58.8%, p = 0.0041). A modest, though statistically significant (p < 0.001), reduction in the number of Ki67 positive cells per field was observed for PLX4720-treated xenografts compared to vehicle controls (77.4% vs. 89.5%).

Figure 5. ¹⁸F-FLT PET predicts lack of response to PLX4720 in HT-29 xenografts

Mice bearing HT-29 xenografts were treated daily with PLX4720 or vehicle. (A) Over a 10 day regimen, no difference in tumor volume was observed between PLX4720-treated and vehicle-treated. Tumor ¹⁸F-FLT PET (B) and phosphorylation of BRAF effectors (C) were similar between vehicle-treated and PLX4720-treated HT-29 cohorts. (D) Representative transverse MALDI IMS pseudocolor maps of ion intensity corresponding to the parent molecular mass of PLX4720 demonstrate minimal signal intensity in vehicle-treated animals. Treated animals demonstrated little PLX4720 ion intensity within tumor tissue (dotted line), yet PLX4720 ion intensity was observed in the bowel.