I would recommend you to do a search for the specific sequence that your antibody recognizes, in this case Dux4, so that you can make a blast with the sequence of the mouse Duxf3 protein and sable if that recognition sequence is present. We have used this procedure for various human antibodies that could be reactive in guinea pigs and we have had satisfactory results.

It all depends on the sequence that the Ab recognizes. A thorough search may help. If not then you may generate an antibody or use expression of mRNAs using in situ hybridization as an additional proof.

Either that, or you just try it out, see what happens.

I have blasted mouse FASTA protein sequence for mouse Duxf3 vs human Dux4 and found only 34% aminoacid similarity you need to have 80%+ similarity for polyclonal antibody to work. If you have a Dux4 monoclonal antibody with known epitope sequence blast p this sequence against mouse Duxf3 if you have an epitome match that it's a very good chance to work otherwise might be a waste of effort.

For publication (let's assume that it works), I strongly suggest adding recombinant or overexpressed Duxf3 positive control. This way by observing the band in the control sample at the predicted size you will be able to convince the reviewers that your results are true.