I don't think that a single word can cover all situations involving the antecedent of a reply. In part, this is due to the nature of written correspondence, as opposed to, say, biological lineage. Whereas in the line from archaeopteryx to chicken (assuming that it's direct), every descendent in the direct line is also an ancestor of the chicken, and the chicken is the descendent of every intervening creature in that direct line, a series of letters or messages and replies lacks any such predictability.
Given the series
Message 1, Reply 1, Message 2, Reply 2, Message 3, Reply 3, Message 4, Reply 4, ...
it is highly unlikely that each reply in the series is a reply not just to the immediately preceding message but to all of the preceding messages.
Even in a particular line of human descent, each parent is linked to a single child, who in turn becomes parent of a single child in that line. Each parent/child in the series has a one-to-one relationship with one other parent and one other child in the line.
In contrast, it's possible for any reply (after the first) in a line of correspondence to be not a response to the immediately preceding message, but a generation-skipping second attempt to answer an earlier question, or a followup to a previous reply in the string of messages, or a message only tangentially related to any of the preceding messages, or an attempt to address remaining issues from multiple previous messages.
What this means is that any word corresponding to reply would have to comprehend such disparate possible meanings as "original message," "second message," "third message," "most recent message," "message of [date]," "one among two or more previous message jointly answered," and "followup to my own reply #n," Under the circumstances, we're better off using multiple words to specify which option we have in mind than trying to come up with a blanket term for all situations.
A second consideration here is that (at least in English) initiating actions aren't especially often defined on the basis of their relation to their consequence. Admittedly some examples do exist:
If B is the effect of A, then A is the cause of B.
If B is a response to A, then A is a stimulus to B.
But far more often (I suspect), our language doesn't go out of its way to establish such reciprocity. Consider these counterexamples:
If B is a reaction to A, then A is a what to B?
If B is a punishment for A, then A is a what for B?
If B is a subject [in the nongovernmental sense] of A, then A is a what of B?
In everyday English, we use many single words to identify results, consequences, or focuses of other things (such as "reaction" as a result of "action," "punishment" as a consequences of "crime," and "subject" as a focus of "study"), but we are somewhat less likely to dedicate single words to identifying a prompt in terms of the response it elicits.