TRPM8 Activation via 3-Iodothyronamine Blunts VEGF-Induced Transactivation of TRPV1 in Human Uveal Melanoma Cells

Abstract

In human uveal melanoma (UM), tumor enlargement is associated with increases in aqueous humor vascular endothelial growth factor-A (VEGF-A) content that induce neovascularization. 3-Iodothyronamine (3-T₁AM), an endogenous thyroid hormone metabolite, activates TRP melastatin 8 (TRPM8), which blunts TRP vanilloid 1 (TRPV1) activation by capsaicin (CAP) in human corneal, conjunctival epithelial cells, and stromal cells. We compare here the effects of TRPM8 activation on VEGF-induced transactivation of TRPV1 in an UM cell line (92.1) with those in normal primary porcine melanocytes (PM) since TRPM8 is upregulated in melanoma. Fluorescence Ca²⁺-imaging and planar patch-clamping characterized functional channel activities. CAP (20 μM) induced Ca²⁺ transients and increased whole-cell currents in both the UM cell line and PM whereas TRPM8 agonists, 100 μM menthol and 20 μM icilin, blunted such responses in the UM cells. VEGF (10 ng/ml) elicited Ca²⁺ transients and augmented whole-cell currents, which were blocked by capsazepine (CPZ; 20 μM) but not by a highly selective TRPM8 blocker, AMTB (20 μM). The VEGF-induced current increases were not augmented by CAP. Both 3-T₁AM (1 μM) and menthol (100 μM) increased the whole-cell currents, whereas 20 μM AMTB blocked them. 3-T₁AM exposure suppressed both VEGF-induced Ca²⁺ transients and increases in underlying whole-cell currents. Taken together, functional TRPM8 upregulation in UM 92.1 cells suggests that TRPM8 is a potential drug target for suppressing VEGF induced increases in neovascularization and UM tumor growth since TRPM8 activation blocked VEGF transactivation of TRPV1.

Keywords: 3-iodothyronamine; Intracellular Ca2+; transient receptor potential melastatin 8; transient receptor potential vanilloid 1 channel; uveal melanoma; vascular endothelial growth factor.

Figure 1

Larger functional TRPM8 expression in UM 92.1 than PM cells. Drug additions were made at the indicated time points (arrows). Data are mean ± SEM of 4–8 experiments. (A) CAP (20 μM) induced an irreversible Ca²⁺ influx (n = 4) whereas non-treated control cells showed a constant Ca²⁺ baseline (n = 4). (B) The same effect could be observed in normal porcine melanocytes but with a time lag (n = 7; controls n = 11). (C) Summary of the experiments with CAP (n = 4–7). The asterisks (*) show significant differences between control and CAP (n = 4; 350s; *p < 0.05; n = 7; 590 s; **p < 0.01; paired tested). The hashtag shows a significant difference of the CAP effect at 350 s between UM 92.1 cells and melanocytes (n = 4-7; 590 s; ^#p < 0.05; unpaired tested). (D–F) Same experiments as shown in (A–C), but with icilin (20 μM) (n = 4–8). The icilin effect was markedly reduced in porcine melanocytes. The asterisks (*) show significant differences between control and icilin (n = 4; 350 s; *p < 0.05; paired tested). The hashtags (#) shows a significant difference of the icilin effect at 350 s and 590 s between UM 92.1 cells and melanocytes (n = 4–8; 350 s; ^###p < 0.005; 590 s; ^##p < 0.01; unpaired tested).

Figure 2

VEGF transactivates TRPV1 channels in UM 92.1 cells. VEGF (10 ng/ml) was added at the indicated time points (arrows). Data are mean ± SEM of 6–17 experiments. (A) Mean trace showing VEGF-induced Ca²⁺ increase (n = 17). (B) Same experiment as shown in (A), but in the presence of CPZ (20 μM). CPZ clearly suppressed the VEGF-induced Ca²⁺ increase (n = 6). (C) Same experiment as shown in (A), but in the presence of BCTC (20 μM). BCTC partially suppressed the VEGF-induced Ca²⁺ increase (n = 6). (D) Same experiment as shown in (A), but in the presence of AMTB (20 μM). AMTB had no effect on the VEGF-induced Ca²⁺ increase (n = 10). (E) Summary of the experiments with VEGF and the TRP channel blockers. The asterisks (*) show significant Ca²⁺ increases with VEGF (n = 17; 300 s; ***p < 0.005; 450 s; *p < 0.05; paired tested). The hashtags (##) indicate statistically significant differences of fluorescence ratios between VEGF with and without the TRP channel blockers CPZ and BCTC, resp. (n = 6–17; 300 s; ^##p < 0.01; unpaired tested). One hashtag (#) indicates a statistically significant difference between CPZ and BCTC effect on VEGF-induced Ca²⁺ increase at 300 s (n = 6; ^#p < 0.05).

Figure 3

Comparison of the effects of CPZ and AMTB on VEGF-induced rises in whole-cell currents in UM 92.1 cells. (A) Time course recording showing the current increases induced by VEGF (10 ng/ml) and after application of 20 μM AMTB. (B) Original traces of VEGF-induced current responses to voltage ramps. Current densities are shown before application (labeled as A), during application of VEGF (labeled as B), and after addition of AMTB (labeled as C). (C–D) Same recordings as shown in (A–B) but with 20 μM CPZ instead of AMTB. CPZ clearly reduced the VEGF-induced whole-cell increases. (E) Summary of patch-clamp experiments with VEGF, AMTB and CPZ. The asterisks (***) indicate statistically significant differences of in- and outward currents with VEGF without and with AMTB (n = 8–40; ***p < 0.005; paired tested). VEGF had no effect in the presence of CPZ.

Figure 4

CAP does not augment increases in whole-cell currents in VEGF treated UM 92.1 cells. (A) Time course recording showing the current peak induced by 10 ng/ml VEGF and current peak after application of 20 μM CAP. (B) Original traces of VEGF- and CAP-induced current responses to voltage ramps. Current densities are shown before application (labeled as A), during application of VEGF (labeled as B), and after addition of CAP (labeled as C). (C) Summary of patch-clamp experiments with VEGF and CAP. The asterisks (**) indicate statistically significant increase with VEGF (n = 7; p < 0.01; paired tested) and unchanged magnitude of currents in the presence of CAP (n = 4; p > 0.05; paired tested). (D) Maximal negative current amplitudes induced by a voltage step from 0 to −60 mV are depicted in percent of control values before application of 10 ng/ml VEGF. VEGF-induced inward currents (n = 7; *p < 0.05) were not increased in the presence of 20 μM CAP (n = 4; p > 0.05). (E) Maximal positive current amplitudes induced by a voltage step from 0 to +130 mV are depicted in percent of control values before application of 10 ng/ml VEGF. VEGF-induced outwardly rectifying currents (n = 7; *p < 0.05) were not increased in the presence of 10 μM CAP (n = 4; p > 0.05).

Figure 5

AMTB suppresses menthol-induced rises in whole-cell currents in UM 92.1 cells. (A) Time course recording showing the current increases induced by menthol (100 μM) and reduction after application of 20 μM AMTB. (B) Original traces of menthol-induced current responses to voltage ramps. Current densities are shown before application (labeled as A), during application of menthol (labeled as B), and after addition of AMTB (labeled as C). (C) Summary of the patch-clamp experiments with menthol and AMTB. The asterisks (*) indicate statistically significant differences of in- and outward currents with menthol without AMTB (n = 8; inward currents; **p < 0.01; outward currents; *p < 0.05; paired tested) and in the presence of AMTB (n = 8; in- and outward currents *p < 0.05; paired tested). (D) Maximal negative current amplitudes induced by a voltage step from 0 to −60 mV are depicted in percent of control values before application of 100 μM menthol. Menthol-induced inward currents (n = 8; **p < 0.01) were clearly suppressed in the presence of 20 μM AMTB (n = 8; **p < 0.01). (E) Maximal positive current amplitudes induced by a voltage step from 0 to +130 mV are depicted in percent of control values before application of 100 μM menthol. Menthol-induced outwardly rectifying currents (n = 8; **p < 0.01) were clearly suppressed in the presence of 20 μM AMTB (n = 8; *p < 0.05).

Figure 6

3-T₁AM elicits increases in whole-cell currents through TRPM8 channels in UM 92.1 cells. (A) Time course recording showing the current increases induced by 3-T₁AM and reduction after application of 20 μM AMTB. (B) Original traces of 3-T₁AM -induced current responses to voltage ramps. Current densities are shown before application (labeled as A), during application of 3-T₁AM (labeled as B), and after addition of AMTB (labeled as C). (C) Summary of patch-clamp experiments with 3-T₁AM and AMTB. The asterisks (**) indicate statistically significant increase with 3-T₁AM (n = 9; p < 0.05 at the minimum; paired tested) and decreases in the presence of AMTB (n = 7; p < 0.05; paired tested).). (D) Maximal negative current amplitudes induced by a voltage step from 0 to −60 mV are depicted in percent of control values before application of 1 μM 3-T₁AM. 3-T₁AM-induced inward currents were clearly suppressed in the presence of 20 μM AMTB (n = 4; *p < 0.05). (E) Maximal positive current amplitudes induced by a voltage step from 0 to +130 mV are depicted in percent of control values before application of 1 μM 3-T₁AM. 3-T₁AM-induced outwardly rectifying currents (n = 5; **p < 0.01) were clearly suppressed in the presence of 20 μM AMTB (n = 4; *p < 0.05).

Figure 7

3-T₁AM elicits increases in Ca²⁺ entry through TRPM8 channels but instead suppresses VEGF-induced Ca²⁺ increases in UM 92.1 cells. 3-T₁AM (5 μM) or VEGF (10 ng/ml) was added at the time points indicated by arrows. Data are mean ± SEM of 4–9 experiments. (A) Mean trace showing 3-T₁AM-induced Ca²⁺ increase (n = 9) whereas non-treated control cells showed a constant Ca²⁺ baseline (n = 13). (B) Same experiment as shown in (A), but in the presence of BCTC (20 μM). BCTC clearly suppressed the 3-T₁AM-induced Ca²⁺ increase (n = 4). (C) Summary of the experiments with 3-T₁AM and BCTC. The asterisks (**) show significant Ca²⁺ increases with 3-T₁AM (n = 9; 350 s; *p < 0.05; 490 s; **p < 0.01; paired tested). The hashtag (#) indicates a statistically significant difference of fluorescence ratios between 3-T₁AM with and without BCTC (n = 6; 490s; ^#p < 0.05; unpaired tested). (D) Mean trace showing VEGF-induced Ca²⁺ increase (n = 6) whereas non-treated control cells showed a constant Ca²⁺ baseline (n = 9). (E) Same experiment as shown in (D), but in the presence of 3-T₁AM (1 μM). 3-T₁AM clearly suppressed the VEGF-induced Ca²⁺ increase (n = 6). (F) Summary of the experiments with VEGF and 3-T₁AM. The asterisk (*) shows a significant Ca²⁺ increase with VEGF (n = 6; 400 s; *p < 0.05; paired tested). The hashtags (##) indicate statistically significant differences of fluorescence ratios between VEGF with and without 3-T₁AM (n = 6; 400 s; ^##p < 0.01; unpaired tested).

Figure 8

3-T₁AM suppresses VEGF-induced rises in whole-cell currents in UM 92.1 cells. (A) Time course recording showing the current increases induced by VEGF (10 ng/ml) and reduction after application of 10 μM 3-T₁AM. (B) Original traces of VEGF-induced current responses to voltage ramps. Current densities are shown before application (labeled as A), during application of VEGF (labeled as B), and after addition of 3-T₁AM (labeled as C). Current densities as function of voltage were derived from the traces shown in panel A. (C) Summary of patch-clamp experiments with VEGF and 3-T₁AM. The asterisks (***) indicate statistically significant increase with VEGF (n = 40; ***p < 0.005; paired tested). The hashtags (##) indicate statistically significant differences of whole-cell in- and outward currents between VEGF with and without 3-T₁AM (n = 7–40; ^##p < 0.01; unpaired tested). (D) Maximal negative current amplitudes induced by a voltage step from 0 to −60 mV are depicted in percent of control values before application of 10 ng/ml VEGF. VEGF-induced inward currents (n = 7; **p < 0.01) were clearly suppressed in the presence of 1 μM 3-T₁AM (n = 7; *p < 0.05). (E) Maximal positive current amplitudes induced by a voltage step from 0 to +130 mV are depicted in percent of control values before application of 10 ng/ml VEGF. VEGF-induced outwardly rectifying currents (n = 7; ***p < 0.005) were clearly suppressed in the presence of 1 μM 3-T₁AM (n = 7; **p < 0.01).

Figure 9

Icilin suppressed VEGF-induced Ca²⁺ increase in UM 92.1 cells. Drugs were added at the indicated time points (arrows). Data are mean ± SEM of 15–85 Ca²⁺ traces in each set of experiments. (A) CAP (10 μM) induced an irreversible Ca²⁺ increase (n = 19). A washout did not reduce the Ca²⁺ levels. Non-treated control cells showed a constant Ca²⁺ baseline (n = 15). (B) The similar Ca²⁺ response pattern could be observed with 10 ng/ml VEGF instead of CAP (n = 85). (C) Same experiment as shown in (B), but in the presence of icilin (10 μM) (n = 65). Icilin partially suppressed the VEGF-induced Ca²⁺ increase (n = 65). Non-treated control cells showed a constant Ca²⁺ baseline in the presence of icilin (n = 36). (D) Summary of the experiments with CAP, VEGF and icilin. The asterisks (***) show significant Ca²⁺ increases with CAP and VEGF (n = 15–85; **p < 0.01 at the minimum; unpaired tested).

Figure 10

Menthol suppressed VEGF-induced Ca²⁺ increase in UM 92.1 cells. Drugs were added at the indicated time points (arrows). Data are mean ± SEM of 15–40 Ca²⁺ traces in each set of experiments. (A) Moderate cooling (≈27 to 18°C) induced a Ca²⁺ increase, which partially recovered (n = 17). Non-treated control cells showed a constant Ca²⁺ baseline (n = 15). (B) TRPM8 activation by menthol (200 μM) led to a Ca²⁺ increase, which is at lower levels compared to moderate cooling (n = 40). (C) Summary of the experiments with cooling and menthol. The asterisks (*) show significant Ca²⁺ increases with moderate cooling and menthol (n = 17 - 40; ***p < 0.005; paired tested). (D) The Mean trace showing VEGF-induced Ca²⁺ increase (n = 25). (E) Same experiment as shown in (D), but in the presence of menthol (200 μM). Menthol clearly suppressed the VEGF-induced Ca²⁺ increase (n = 32). (F) Summary of the experiments with VEGF and menthol. The asterisks (*) show significant Ca²⁺ increases with VEGF (n = 25; ***p < 0.005; paired tested). The hashtags (###) show significant Ca²⁺ decreases in the presence of menthol (n = 32; ^###p < 0.005; unpaired tested).

Figure 11

3-T₁AM modulation of TRPV1 is associated with cannabinoid receptor type 1 activity. Drugs were added at the indicated time points (arrows). Data are mean ± SEM of 13–53 Ca²⁺ traces in each set of experiments. (A) WIN 55,212-2 (10 μM) induced a Ca²⁺ increase, (n = 27). Non-treated control cells showed a constant Ca²⁺ baseline (n = 34). (B) Ca²⁺ free condition reduced the intracellular Ca²⁺ level (baseline) and extracellular application of WIN 55,212-2 (10 μM) strongly increased the intracellular Ca²⁺ level (n = 53). (C) Summary of the experiments with WIN 55,212-2 with and without external Ca²⁺. The asterisks (***) show significant Ca²⁺ increases with WIN 55,212-2 (n = 27 - 53; ***p < 0.005; paired tested). The hashtags (###) show significant stronger Ca²⁺ increase under external Ca²⁺ free conditions (n = 53; ^###p < 0.005; unpaired tested). (D) Mean trace showing 3-T₁AM-induced Ca²⁺ increase (n = 13) whereas non-treated control cells showed a constant Ca²⁺ baseline (n = 29). (E) Same experiment as shown in (D), but in the presence of AM251 (10 μM) (n = 46). The 3-T₁AM-induced Ca²⁺ increase is at higher levels compared to the effect without the CB1 blocker. (F) Same experiment as shown in (B), but with 1 μM 3-T₁AM instead of WIN 55,212-2. 3-T₁AM did not change the intracellular Ca²⁺ concentration. (G) Summary of the experiments with 3-T₁AM with and without AM251 or external Ca²⁺. The asterisks (***) show significant Ca²⁺ increases with 3-T₁AM (n = 13–46; ***p < 0.005; paired tested). The hashtags (###) show significant difference of the 3-T₁AM-induced Ca²⁺increase with and without AM251 (n = 13–46; ^###p < 0.005; unpaired tested).

Figure 12

Suggested Ca²⁺ signal transduction model accounting for how TRPM8 activation affects receptor-linked signaling pathways (Mergler et al., ; Khajavi et al., 2017). Ca²⁺ channels such as TRPs of the TRPV1 subtype (capsaicin receptor) can be selectively activated by heat (>43°C) or capsaicin and blocked by CPZ (Figure 1A) (Mergler et al., 2014). VEGF-A activating VEGFR-1/VEGFR-2 can also activate TRPV1 (Figures 2, 3). The TRPM8 subtype (menthol receptor) can be selectively activated by cold (23–28°C), menthol, icilin or 3-T₁AM and blocked by BCTC/AMTB (Figures 1D, Figures 5, 6) (Mergler et al., ; Khajavi et al., 2017). Notably, a G-protein coupled receptor (GPCR) coupled to Gi/o proteins could be activated by 3-T₁AM (Dinter et al., ; Khajavi et al., ; Schanze et al., 2017). 3-T₁AM may also directly activate TRPM8 by a GPCR-independent mechanism (↑[Ca²⁺]_i]) (Khajavi et al., 2017). If TRPM8 is activated by 3-T₁AM, 3-T₁AM suppresses VEGRF via TRPV1 (Figures 7, 8). Notably, 3-T₁AM may also directly suppress TRPV1 via VEGFR by a GPCR-independent mechanism (↓[Ca²⁺]_i]) (Figure 8) Menthol and icilin mimic the 3-T₁AM effect and suppresses increases in TRPV1 activity by VEGF (Figures 9, 10).