About proving the logical syntactic axiom 3's reverse by using, (deducing from), axiom1 and 2.

Question

The syntactic axioms here are

1. $\forall p,q \in L : p\Rightarrow (q \Rightarrow p)$

2. $\forall p,q,r \in L : (p\Rightarrow (q \Rightarrow r))\Rightarrow ( (p\Rightarrow q) \Rightarrow (p \Rightarrow r) )$

3. $ \forall p \in L : \neg \neg p \Rightarrow p$

We know that axioms 3 independent from 1 and 2, however, its reverse

$ \forall p \in L : p\Rightarrow \neg \neg p $ is not.

So exactly how to show the above reverse by using axiom1 and 2?

I firstly change it into a more original form

$a\Rightarrow ((a\Rightarrow\bot)\Rightarrow\bot)$.

By using the Completeness Theorem, such a question will be trivial, so if not using it but simply use Deduction Lemma, I wonder how to show this. I tried a long time but didn't succeed so I wonder if my method is not correct.