A starting point is the smallest sample return probe ever built, the Luna 24.
Massing 5300 kg in lunar orbit it was pretty much bare bones. To see where we can improve this, we must first split the mission it its separate parts. This follows the standard procedure of doing the mission backwards. First, I would not choose ISS as the target, as braking into orbit and then rendezvous and dock with it is a complicated task. To simply hit the Earth is much easier.
Of the 514kg for the return stage of the Luna 24, around 300kg was propellant. Perhaps a slightly more efficient engine exist today, but the technology used now for landing and ascent would still be hypergolic propellants. That is the most important limit for scaling.
The re-entry capsule was only 34kg, but you might still be able to shave off a few kilograms. The main savings are in the remaining 180kg though, including electrical systems, control systems, the engine and the propellant tanks. With the miniaturization of electronic equipment since the seventies, and some new lighter materials, you may be able to squeeze everything required into a dry mass of 100kg. That is about 220kg at the lunar surface, roughly halved. A similar miniaturization of the descent stage and drilling equipment yields a spacecraft of about 2 metric tonnes in Lunar orbit.
To transport the spacecraft to the Moon, the minimal solution is a spiralling ion-craft. That is within current technology, take for instance the engine powering Dawn. At the cost of a long transfer time, the total mass in Earth orbit is only going to be around 3 metric tonnes minimum. However, considering the relatively high drag at the altitude of the ISS, the craft must start from a higher orbit.
Additionally, it is more effective to combine multiple related goals, like a lunar rover, and on-site experiments, than to launch multiple minimal missions with high risk of failure.