Proto-Indo-European was spoken in mid-east Russia, directly above the Caucuses. Their ancestors were probably the Yamna culture, located further east, closer to Mongolia or south central Russia. Just as I demonstrated in my Urheimat post about Europe is true for Asia: a stretch of terrain from Northeast Asia to Northwest Europe below the permafrost was populated by dozens of language families, if not more. In fact, there is no reason to doubt that the same stretch functioned as a long route of inter-tribal contact and trade.
There are at least three linguistic hints of early trade. The trick, however, is that dating the entry of the words is a tricky ordeal, so the question of "When did Lexeme A enter Language 1?" is frustrating. In most cases, we can say that a word may have entered into a language before a certain date but not after. So we need a disclaimer that these hints may be several thousand years older than each other.
The word for sulfur has resisted etymology in a good many language families. Semitic, Germanic, Italic, Slavic (in this case, the Italic, Slavic and Germanic reflexes are remarkably similar but not directly related), Turkic (in Turkish and Mongolian), etc... Further, in many of the languages, the reflex may have cognates in unrelated languages. This is proof positive of language contact, and strong evidence of trade. But what is more exciting is that if both cognates come are loans, then they must have come from a third-party language: an ancient, disappeared language (!) Suddenly historical linguistics tells us things before the first ancient writings and outside the soil of archaeological digs.
2. Cannibas and 3. Hemp
These two examples are fairly convoluted, so I'll try to give a simple explanation. Both words trace back to separate reflexes. hemp < Proto-Germanic *hanipa and cannibas < Proto-Germanic *kanib-. There's an obvious relationship between the two. First, there's a mechanical relationship, as hemp is made from the cannibas plant. Second, the linguistic relationship is tantalizingly close; an /h/ ~ /k/ and /p/ ~ /b/ connection is exceedingly safe to make. However, such a change between the two consonants has no morphological or phonological foundation in German. Because there is no Proto-Indo-European etymology for the word, and because there are cognates in Turkic and other North Eurasian languages, a trade route may be suggested.