Robert T. answered 10/15/19
Masters degree in biology with 4 years of teaching experience
Since both of these instances are examples of dominant mutations, just ignore that aspect, for now. What you should focus on, instead, is that the first case is a loss-of-function mutation that leads to haploinsuffeciency, meaning that the gene product (let's assume it's a protein, for simplicity) must be at a high enough level, a "critical mass", if you will, in order to have correct phenotype. In short, this instance is a result of missing one functional allele where 2 functional alleles are required.
In the second instance, this is a "classic" gain-of-function mutation. These are almost always bad as the new "function" doesn't fit in with the intracellular signaling "ecosystem" and tends contribute to vast disruptions in cell function. In this case, the mutation results in one functional allele, and one allele that has new function(s) that is/are "foreign" to the normal operation of the cell.
So, with current clinical trials for gene therapy, only by adding an additional functional copy can rescue the wild type phenotype. One normal allele, one loss-of-function allele, and one introduced normal allele should be enough to "get over the hump" and make enough of the wild type gene product to have wild type phenotype, so I believe only A is the answer. Even if you add several of the wild-type copies in a gain-of-function mutation, you still have an allele that makes a product that has a new phenotype. So, I don't think it can be B.
You might be able to play devils advocate and suggest you introduce extra copies of a gene that inhibits the gain-of-function allele, but that's far more complicated (you may need to introduce several genes for this to work, and we don't have the technology to do this reliably in humans). Hope that helps!