Towards a Bootstrap approach to higher orders of epsilon expansion