They queried publications dealing with the lncRNA that studied whether it was functional through over-expression or knockdown. Cells do something measurable in their normal state, that can be observed through microscopy, qPCR, microarray, etc. You use the data as a control which to compare experimental results against.
For an over-expression study, for example, you can transfect your cells with a vector that will transcribe a lot of your ncRNA. Whether the transfection need be stable or transient will depend on a number of other factors. So the methods you used to collect data as a control, apply these to the transfected group. If there are statistically significant changes to anything when you have more of the ncRNA, you can say in part that your ncRNA is functional. It's doing something observable.
The opposite is true for knockdown. You can maybe use CRISPR to do the job but lentiviral transduction with say a doxycycline-inducible promoter is among other possibilities. The goal now is to make the gene, or the expression of the gene go away. You do the same experiment, and see if an absence of the ncRNA does anything to the cell.
Now in the database you're looking at, expression is being used to show you in what tissues has BX118339 been detected? It basically says BX118339 is expressed in testis and the fetal brain.