In the meantime, I found a workaround with FFMPEG :
1/ I export from Resolve the video silent (without audio) as an .MXF OP1a AVC-Intra 100 1080i25 (every flavor I need)
2/ I export from Resolve the audio as stereo .WAV (48 Khz 24 bits)
3/ I rewrap without converting/re-compressing via FFMPEG :
ffmpeg -i VideoSilentFile.MXF -i AudioStereoWaveFile.wav -map 0:0 -map 1:a -vcodec copy -acodec copy VideoWithStereoAudio.MXF
And voilà : an .MXF file with only one stereo audio track containing two audio channels 🙂