It's not quite vice grips; I had to dig into the paper to understand it. Their specific DTAL protocol provided "moderate mechanical loading" of "2 N at 2 Hz for 300 cycles, twice/week, [for] 3 months". For reference, a 200g stationary object exerts that much force on a surface due to gravity. With 300 cycles at 2Hz, the exercises lasted for 150 seconds, or 2.5 minutes, and the exercises occurred twice a week.

Note, however, that their experimental model section says they applied axial force (think columns, not joists) to the tibia using an ElectroForce 3300 (https://www.tainstruments.com/3330-system/) while the mice were under anesthesia. From this, it seems like they knocked out the mice and placed their legs underneath a mechanical ram to apply axial compressive forces.

Because people are typically awake when exercising, given that muscles are designed to carry forces around the skeleton, I imagine, even after scaling from a mice to a human purely by sizing, the human force equivalent will be many times higher.

If you're looking to try this out for yourself, the seated leg press will most closely simulate the study's method. Figure out what "moderate" loading is for you, and shoot for 3 sets x 10 reps at least twice a week. Overall, choose skeletally-compressive exercises where you're "pushing," such as weighted squats (e.g. barbell squats) and any type of "press" exercise like the leg press, chest press, military press, and even cycling (you push your pedals). For the purposes of this study, avoid exercises with any pulling motion, including pull-ups, pull-downs, rows, and swimming.

From a force perspective, running has some compressive elements, but the load comes from gravity alone, and good runners minimize compressive forces. Since running therefore is primarily a pulling exercise, I can sort of see what they're getting at.