Thanks for your testing so far. I have been doing research on this, there is a long way to go. I may know what I am doing in a few years. Please let me know if I have misunderstood anything. My brother recently gave me his old Nikon D70 so I can try imaging. My experiments are still a few weeks off before I can go outside to play.
What I brought up in the other thread that started this was best ISO for different cameras. There is a point above which the ISO setting will reduce your dynamic range. There is information available, by the research of other more knowledgeable people than me, as to what the optimum ISO setting for different camera types is for best dynamic range.
Since a DSLR chip has a fixed sensitivity, the ISO setting acts as an amplifier of the chips analog signal before it is converted to digital. Increasing the ISO does not change the signal the chip has, it just multiplies it before being converted to digital.
Your images as taken do show increased brightness, due to the sensor data being multiplied by the ISO changes. For your experiment to show what I mentioned in the other thread about ISO setting for best dynamic range, your exposures at different ISO settings would also require different exposure times. 800 ISO would need to be exposed twice as long as 1600, and so on. This would show if the image was starting to loose it's dynamic range.
Now it seems to me, part of the issue also depends on how well your telescope tracks. An exposure time beyond the setups tracking ability, or objects brightness, will of course be of no use. To optimize image data is quite a balancing act.
What I found so far applies to RAW images. Data taken before the camera software alters the image. I believe Cannon cameras can do this easily, while an older Nikon like my D70 requires you to turn the camera off after the data is acquired to stop the software processing of the image data.
Of course, from what I have read, stacking can work wonders. Multiple images that are stacked at low light levels can increase dynamic range while reducing noise. Then it seems I have even more to learn with darks, lights, and processing the images after acquiring them.