Parsing XML with a non-syntax aware tool like sed is a bad idea. The solution would break is there is a slightest change in the XML formatting. Use xmlstarlet which is syntax aware tool for editing files.
Since your file is syntactically valid also, it is pretty straightforward to solve your problem. Just do
xmlstarlet edit --update '//NikuDataBus/LookupQuery/Filter' --value 'lookup' input.xml
To make the edit of the file in-place just like sed -i, use the -L flag to the solution if your version supports it or use a temporary file (> temp && mv temp input.xml)
xmlstarlet edit -L --update '//NikuDataBus/LookupQuery/Filter' --value 'lookup' input.xml
Or you could go one step further and do the replace only if the Filter tag contains test which you can do as
xmlstarlet edit --update '//NikuDataBus/LookupQuery/Filter[contains(.,"test")]' --value 'lookup' input.xml
Tested on Centos 7 on the following version.
$ xmlstarlet --version
compiled against libxml2 2.9.1, linked with 20901
compiled against libxslt 1.1.28, linked with 10128
Thanks to one of the useful comments from ever Perl expert - Sundeep, you could see the list of paths that could be traversed on which you can apply operations with xmlstarlet
$ xmlstarlet el input.xml
NikuDataBus
NikuDataBus/Header
NikuDataBus/LookupQuery
NikuDataBus/LookupQuery/Filter
{}button orctrl+kto format the data for better viewing... also, use parsers like xmlstarlet instead of sed for editing xml data