Critical role of transient activity of MT1-MMP for ECM degradation in invadopodia

Abstract

Focal degradation of extracellular matrix (ECM) is the first step in the invasion of cancer cells. MT1-MMP is a potent membrane proteinase employed by aggressive cancer cells. In our previous study, we reported that MT1-MMP was preferentially located at membrane protrusions called invadopodia, where MT1-MMP underwent quick turnover. Our computer simulation and experiments showed that this quick turnover was essential for the degradation of ECM at invadopodia (Hoshino, D., et al., (2012) PLoS Comp. Biol., 8: e1002479). Here we report on characterization and analysis of the ECM-degrading activity of MT1-MMP, aiming at elucidating a possible reason for its repetitive insertion in the ECM degradation. First, in our computational model, we found a very narrow transient peak in the activity of MT1-MMP followed by steady state activity. This transient activity was due to the inhibition by TIMP-2, and the steady state activity of MT1-MMP decreased dramatically at higher TIMP-2 concentrations. Second, we evaluated the role of the narrow transient activity in the ECM degradation. When the transient activity was forcibly suppressed in computer simulations, the ECM degradation was heavily suppressed, indicating the essential role of this transient peak in the ECM degradation. Third, we compared continuous and pulsatile turnover of MT1-MMP in the ECM degradation at invadopodia. The pulsatile insertion showed basically consistent results with the continuous insertion in the ECM degradation, and the ECM degrading efficacy depended heavily on the transient activity of MT1-MMP in both models. Unexpectedly, however, low-frequency/high-concentration insertion of MT1-MMP was more effective in ECM degradation than high-frequency/low-concentration pulsatile insertion even if the time-averaged amount of inserted MT1-MMP was the same. The present analysis and characterization of ECM degradation by MT1-MMP together with our previous report indicate a dynamic nature of MT1-MMP at invadopodia and the importance of its transient peak in the degradation of the ECM.

Figure 1. Critical role of the MT1-MMP turnover at invadopodia and a sharp transient peak in MT1-MMP.

(A) Degradation of ECM by MT1-MMP is inhibited by the application of Dynasore blocking vesicular turnover. Green: ECM (fibronectin), Red: actin, Black dots: focal degradation of ECM. (B) Model used in the present study is schematically shown. MT1-MMP in the invadopodial membrane is inserted and internalized by vesicular trafficking, MT1-MMP forms dimer, ternary complex with TIMP-2 and proMMP-2. ProMMP-2 is activated by TIMP-2-free MT1-MMP. ECM is degraded by active complexes of MT1-MMP and MMP-2. (C) Simulation revealed the existence of a sharp transient peak at the initial phase of the formation of active MT1-MMP complexes (M14a). The height of the transient peak depends on the TIMP-2 concentration (inset). (D) Steady state level of M14a (M14a_steady) is plotted as a function of TIMP-2 concentration. There is an abrupt decrease in M14a_steady in the region of TIMP-2 from 80 nM to 120 nM.

Figure 2. Rapid formation of inactive complexes creates a sharp transient activity of MT1-MMP.

(A)Time courses of all complexes comprising M14a, M14, M14.M14, M14.M14.T2, and M14.M14.T2.M2, are shown. The concentration of M14 decreased rapidly in monotonic fashion, concentrations of M14.M14 and M14.M14.T2.M2 resemble biphasic time courses. The concentration of M14.M14.T2 also shows a biphasic time course but is present in negligibly small concentration. (B) Time courses of all inactive complexes of MT1-MMP are shown. Concentrations of M14.T2, M14.T2.M2, and M14.T2.M14.T2.M2 resemble biphasic time courses, while the concentration of M14.T2.M2.M14.T2.M2 increases monotonically until reaching a steady plateau level. The concentration of M14.T2.M2 is also present at a significant level at steady state. The numbers at the left of each complex indicate the complex number. (C) The change in the time course by path-deletion. If pathways leading to the complexes of M14.T2.M2.M14.T2.M2 (deletion 1) and M14.T2.M2 (deletion 2) are deleted, M14a_steady is considerably higher than the concentration in the presence of both pathways (control). If we delete all pathways to all of the inactive complexes (deletions 1 through 5), there is an additional increase in M14a_steady. (D) M14a_steady progressively increases together with the increase in the number of deleted complexes.

Figure 3. Changes in the M14a_steady according to the change in the turnover rate.

MT1-MMP undergoes turnover based on our experimental observation , and the effect of the change in the turnover rate on M14a_steady is investigated by simulations. (A) In the presence of the turnover, M14a_steady is considerably higher at TIMP-2 concentrations higher than 80 nM (red continuous line) in comparison to the level in the absence of turnover (black line). If we reduce the turnover rate progressively, M14a_steady reaches the level seen in the absence of turnover (red broken lines). (B) Active complexes increase together with the increase in the turnover rate. Among them, M14 displays the most significant increase, followed by M14.M14.

Figure 4. The peak transient activity of MT1-MMP is critical to the effective ECM degradation.

(A) In the presence of the sharp transient activity in M14a, ECM is degraded effectively depending on the TIMP-2 concentration. τ_H is defined as the time at which half of ECM is degraded (top panel). If the transient activity is eliminated computationally, the ECM degradation is greatly delayed (bottom panel). (B) The reduction in the efficacy of ECM degradation by the elimination of the sharp transient peak is clearly shown. At TIMP-2 concentrations lower than 40 nM, only a small change in τ_H is seen. At 80 nM of TIMP-2, however, τ_H increases dramatically in the absence of the sharp transient activity. The difference in τ_H is more than two orders of magnitude at TIMP-2 of 500 nM.

Figure 5. Pulsatile insertion of MT1-MMP.

(A) Time courses of M14a in the pulsatile insertion model of MT1-MMP. There are two different intervals in pulsatile insertion of MT1-MMP, which correspond to pools X and D. This results in the double zigzag lines with short (2.6 sec for pool D) and long (25.9 sec for pool X) intervals (top panel). If the intervals of the insertion of MT1-MMP in pools X and D are random instead of regular, M14a follows a considerably random time course (bottom panel). (B) Comparison of time courses for ECM degradation between regular (continuous lines) and random (broken lines) pulsatile insertion intervals. The differences in the time courses are small. (C) If we plot the time courses of complexes of active MT1-MMP and ECM (M14a.ECM) at TIMP-2 of 100 nM, its randomness in the random pulsatile insertion interval is much reduced in comparison to the time course of M14a (bottom panel in A). If the integrated values of M14a.ECM for regular and random pulsatile insertion, which is the direct measure of degraded ECM, are plotted, the curves resemble smooth lines for random pulsatile insertion (thin red line) and regular pulsatile insertion (thick red lines). There is only a small difference between the two.

Figure 6. The effect of MT1-MMP concentration at each insertion while keeping time-averaged insertion amount unchanged.

(A)Two examples of pulsatile insertion of MT1-MMP. The upper example shows small but frequent insertions, while the lower example shows three-times larger insertions but with one-third of the insertion frequency. In both cases, the time-averaged insertion amount of MT1-MMP is unchanged. The problem is whether these two regimens result in the same time courses of ECM degradation. (B) There is no change in τ_H by the fold-increase in frequency/concentration of pulsatile insertion from 100/0.01 to 0.2/5, while it decreases with the change from 0.1/10 to 0.01/100. This indicates that ECM degradation proceeds faster in cases of higher MT1-MMP concentration in a single vesicle with longer insertion intervals, even if the time-averaged amount of MT1-MMP is not changed. The inset shows the time courses of ECM degradation. At the control condition (1/1), the frequency/concentration are (1/25.9 sec)/3 nM and (1/2.6 sec)/7 nM for pools X and D, respectively. (C) There is a positive correlation between the initial rate of ECM degradation and the fold-change in insertion frequency/concentration. It is expected that the ECM remaining at the next insertion is the same even if the frequency/concentration regimen is different (inset). (D) At a higher frequency/smaller concentration regimen, the ECM concentration remaining at the next insertion is almost unchanged with the change in the regimen (black and blue lines). However, at lower frequency/higher concentration regimens, there are significant differences in the remaining ECM at the time of later insertions (t_a and t_b). This results in the larger value of (inserted MT1-MMP)/(remaining ECM), and a larger amount of ECM is degraded with the same amount of inserted MT1-MMP at later insertions. Thus, the difference in the remaining ECM (d_a and d_b) increases at later insertions.

Figure 7. Spatio-temporal simulation of ECM degradation in the presence and absence of the sharp transient activity.

(A) ECM degradation in the 3D model. The 3D space was cuboid with axes of 5 µm×5 µm and 3 µm (inset). The space was divided into 51×51×1 compartments. One surface of the space was assumed to be the ventral surface of a cancer cell, and ECM was present all along the space. We set 49 compartments at the center as an invadopodium (shown in red in the inset) whose diameter was about 0.9 µm. The time courses for continuous (continuous lines) and regular pulsatile (broken lines) insertion in the presence (red lines) and absence (black lines) of the sharp transient activity at TIMP-2 of 200 nM are shown. In both cases, the sharp transient activity in M14a plays a critical role, and virtually no ECM degradation is seen in its absence. (B) Line-scans showing spatio-temporal change in the ECM degradations. Spatial profile was taken along a white line. There is virtually no difference between continuous and pulsatile insertion in the spatio-temporal dynamics of ECM degradation, but there is considerable difference between in the present and absence of transient M14a activity. No appreciable ECM degradation is seen when transient activity of M14a is eliminated.

Figure 8. The dynamics of MT1-MMP for degrading ECM at invadopodia.

There are two docking sites in the invadopodial membrane to which MT1-MMPs are inserted (pools D and X). The turnover rates of MT1-MMP in pools D and X are 26.0 sec and 259 sec, respectively, and MT1-MMP to these pools is transported to these pools by vesicular trafficking. MT1-MMP inserted into these pools is subjected to ECM degradation and also to inhibition by the endogenous inhibitor TIMP-2. The present study has shown a quick inhibition by TIMP-2, and hence the half-life for newly inserted MT1-MMP to maintain ECM-degrading activity is only about 4 sec depending on the TIMP-2 concentration. This short life-time is compensated by the quick turnover of MT1-MMP resulting in the effective degradation of ECM at invadopodia.