Custom estimator for gridsearch

I have X_train with shape(100,6) and y_train with shape(100,3) and I want to use SVR to predict y_test. Unfortunately, SVR does not support many dimensional target variable, so I implemented my class:

from sklearn.base import BaseEstimator, clone

class VectorRegression(BaseEstimator):
    def __init__(self, estimator, **parameters):
        self.estimator = estimator(**parameters)

    def fit(self, X, y):
        n, m = y.shape
        # Fit a separate regressor for each column of y
        self.estimators_ = [clone(self.estimator).fit(X, y[:, i]) for i in range(m)]
        return self

    def predict(self, X):
        # Join regressors' predictions
        res = [est.predict(X)[:, np.newaxis] for est in self.estimators_]
        return np.hstack(res)

Now I want to find best parameters for SVR using GridSearchCV, but I do not know how to assign this parameters to my VectorRegression estimator, because it is already declared when plugged into GridSearch.

Tell me please how to change the interface to make VectorRegression be able plugged into GridSearch.