How certain are you about the gene annotation? In mammals, splice variants are how we get more transcripts than genes (and by a LOT). So, the second "intron" might be a coding region for some contexts.

A lot of splicing relies on 'context', in ways we don't necessarily understand very well. For example, exon-skipping (forced exclusion of specific exons via antisense oligonucleotides) works usually by blocking splice motifs, so if you mask the splice acceptor site of an exon, the spliceosome skips over that exon and splices to the next viable acceptor.

Weirdly, this works even when targeting splice DONOR sites: you'd think this would just kill splicing entirely (exon 1 spliced to 2, then 2 to 3, then block 3 and...no donor left, transcript left incomplete) but instead the spliceosome somehow spots that exon 3 won't be downstream splicing competent and skips exon 3 anyway.

You can also get non-sequential splicing (2, 3, 4 all spliced, THEN 5, 6, 7, THEN 1 spliced to 234, then 1234 spliced to 567).

And introns themselves contain splicing enhancers and suppressors.

So it's entirely possible that forcing retention of intron 1 will have broader consequences: for example, if splicing of intron 2 requires removal of splice suppressor sequences in intron 1, then it'll be preferentially retained.