I'm not sure a mismatch between slope and correlation is necessarily a problem. Note that the slope doesn't equal the correlation with Pearson's r either, except when the variables are standardized to have equal variance.
More generally, no, I don't think its too meaningful to put a line on the scatterplot to represent a Spearman correlation. Such a line implicitly reflects a linear scaling within X and within Y, whereas Spearman only assumes ordinality within each of X and Y.
