This depends on what RNA they are looking at. A typical mammalian cell contains about 10-30 pg of total RNA, most of this is rRNA and tRNA, leaving about 5% mRNA. Of this there are approximately 12,000 transcripts on average 1.5 kb in length, with abundant mRNAs comprising about 3% (about 400 molecules) and rare mRNAs about 0.1% (about 10 molecules).
So for some experiments you can get away with single cells, however I would recommend 30+ cells and do one-step RT-PCR using specific primers as you can use all of the sample.