Inconsistent results between sjPlot and Anova()when including interaction terms in lmer() in R

Question

I am using this R package "sjPlot" to plot the forest-plot of standardized beta values. My model includes interaction terms

m1.3.5=lmer(Q3~1+Q4+gender+gcOUT*gender+maxXcorr*gender+(1+Q4+gcOUT*gender+maxXcorr*gender|sub))

The "gender" variable is binary, and the others are continuous variables.

The results of Anova (car package)

Anova(m1.3.5)
Analysis of Deviance Table (Type II Wald chisquare tests)

Response: Q3
                   Chisq Df Pr(>Chisq)    
Q4              228.6415  1  < 2.2e-16 ***
gender            0.4101  1   0.521927    
gcOUT             7.8720  1   0.005021 ** 
maxXcorr          6.2570  1   0.012370 *  
gender:gcOUT      2.3550  1   0.124884    
gender:maxXcorr   0.5016  1   0.478812  

Here is the plot:

plot_model(m1.3.5, type = "std",se=TRUE,line.size=1.5,dot.size=4,show.values = TRUE)

The plot and the Anova() results seems to be quite different. For example, the gcOUT effect is significant in Anova. However, the SE of the plot of that variable crossed 0, and it does not have *, suggesting it's not significant.

May I ask why the results of these 2 were inconsistent?

Note that if without interaction variables, the results of these two are the same.

Thank you all for your great help in advance :)