Not being a specialist on delivery, never heard of intrathecal delivery of anything but drugs or plasmid codding for Cytokine/hormones. Remember that your target is in nuclear. Mammalian cells do not have efficient DNA capture, may be few cells around place of injection will be transformed. More common method of delivery in mice is all sorts of viral delivery. If you are targeting neural tissue, - HSV and some subtypes of AAV should work.
Agree with Dmitry. However, we are trying delivery of plasmids in a calcium solution in delivery to both muscle and bone. We have generated a small number (2) of tumors but have not yet confirmed if it is truly from the Indel we were trying to create with CRISPR.