This is a great question. In fact it is THE question that helped draw me into folklore studies more than a half century ago.

There are several issues to consider here. I avoid the term "religion" to describe what is going on here because the term conjures an image for the modern mind that doesn't really apply very well. We tend to think of religions as something filled with doctrine, dogma, and a professional, organized priesthood together with a written text holding everything together. Of course, that image is undermined by the fluidity of everything we describe today as "religions." Even religions change from generation to generation and place to place.

Greek, Nordic, and Vedic pantheons (whatever that means) find imperfect documentation in stories that are preserved from historic (often ancient) times. These stories are reflections of and reinforced belief systems that were themselves fluid. In their earliest manifestations, there were no written texts nor was there dogma/doctrine. And what priests existed may have been professional with a lower case "p", but they did not belong necessarily to a priesthood in the modern sense of the word.

So, we are trying to observe what is basically a variety prehistoric belief systems through their expression in stories, written down at a later stage in the development (or in the case of Nordic cultures, after the extinction) of those belief systems. We must keep in mind that while religions change over time and geography, so too do belief systems, particularly when they are untethered to a written sacred text and to an organized professional priesthood. We can talk about a Norse pantheon, but when we try to pin down the when and where of it, things become problematic: some cohesion likely tied the whole thing together, but variation ruled this domain of Northern European cultures before the conversion.

But let's say for sake of argument that we can reconstruct the quintessential Green, Norse, and Vedic pantheons respectively, and that we can use these to triangulate on "the" ancient Indo-European pantheon/belief system. What we are chasing, here, is an idealized moment when that culture was the common ancestor of all of these offspring. The problem we encounter is that "that culture" never really existed. Given that belief systems and their accompanying stories are fluid, whatever existed in ancient Indo-European prehistory was certainly fluid, spread over at least some geography, and traveling through time, changing all along the way.

Efforts to reconstruct that early Indo-European language is equally problematic and reflects what we face when it comes to trying to understand early story and belief. We can theorize and reconstruct what the early Indo-European root of a word might have been, but that theoretical word may have only been spoken for a generation and only in one valley. Language changes over space and time just like belief systems and stories.

Efforts to reconstruct an "ancient Indo-European religion" are flawed because whatever that cultural institution was like among the earliest culture(s) that spoke Indo-European, it was in flux. More importantly, your question, "what did that religion originate from" betrays the real problem. Even if we could triangulate on that early "ancient Indo-European religion" we would inevitably ask the question - since belief systems and their stories are always in flux, what went before? The answer to that question is maddening.

We can imagine an answer, however, with a thought experiment. Let's imagine we have a time machine (and enough antibiotics to survive our journey!). We travel back to the Eurasian Steppes and document an ancient Indo-European belief system. We then travel to the next valley, and the next, documenting variation. We have arrived at an understanding of the range of possibilities at that time, and we have done a good job at documenting the loosely bound beliefs and stories that represent something of a cultural cohesion.

We then ask your question - but what went before? So we use our time machine to repeat the exercise, traveling several hundred years back at a time, tracking the ancestors or our first ethnographic study and documenting their belief systems and stories, and we begin to build an image of how things mutated and drifted over time. But this still doesn't satisfy our curiosity because we want to know what went before, so we repeat the exercise, jumping several hundred years at a time to earlier and still earlier times.

After repeating this exercise several dozen times, we develop a better understanding of the range of possibilities and how this belief system drifted and mutated, but what never emerges - and never can emerge - is a definitive "earlier religion" that spawned everything else. We can begin to see how some stories diffused in from elsewhere and were adapted as they were adopted, and we can see how other beliefs and their stories changed over time and space. The exercise would help us to understand a great deal about the earliest cultures - and more importantly culture changes - or the ancestorial speakers of Indo-European, but we would fail to arrive at the "definitive" religion.

Secondary source books on ancient mythologies can be deceptive in the way they lull readers into thinking that they are having some ability to merge into an ancient religion. These cultural institutions were not locked in stone - or print - until fairly late in the game. Change and variation are key to understanding what they were really about.

edit: thanks for the gold! Much appreciated!