Following up on my previous question about a latin word for "dunes", how would the Romans have adopted and latinized a word of Germanic origin like "dune"? I found a few examples of similar sounding words like dunum and dunus. I think the former might be a Gallic or Celtic town name that was latinized, and the later might be a mispelling of dumus, though.
It depends on the exact word they'd be trying to borrow. Proto-Germanic was spoken in the first centuries CE and the Proto-Germanic word reflected as English dune is reconstructed as *dūnǭ or *dūnaz; the latter would be a straightforward dūnus (compare e.g. rattus 'rat' < PGmc *rattaz), the former likely dūnō (third declension, like sāpō 'soap' < PGmc *saipǭ), conceivably dūna (especially if the speaker is educated and fluent enough to notice *dūnǭ is feminine).
Going earlier, the Proto-Germanic is probably borrowed from Gaulish *dūnon or its predecessor, Proto-Celtic *dūnom, which would be borrowed into Latin as dūnum, as actually happened in the compound Lugdūnum < Gaulish *Lugudūnon (mentioned in your link).
Going later, *dūnǭ became dūna in the West Germanic languages, which is straightforward to borrow (dūna), and then after that the Germanic languages began to lose their endings. In Middle Dutch dūna weakened to dūne /dyːnə/, and it's this word that gave the Romance languages their words, all of which, like the Dutch, are feminine: French dune, Spanish, Portuguese, Italian duna, Romanian dună (presumably not a direct borrowing). Dutch duin, like English down (dune is a borrowing from French) and Swedish and Icelandic dyn, but unlike German and German Low Saxon Düne, eventually lost the ending entirely, in which case Latin seems to pick a first or second declension ending more or less at random. Since related words like sabulō, collis, and mōns are all masculine I'd probably go with dūnus at that point (possibly diūnus if borrowed from English dune, maybe deunus if from Dutch duin), but there's no firm rule.
In general: if the word they're trying to borrow has an ending that looks like an existing Latin nominative ending (which can happen surprisingly often: most of Latin's neighbours were Indo-European, and they resembled each other more closely at the time than they do today; on the other hand, Latin did lack a lot of phonemes its neighbours had, so genuine correspondences were also often missed), that's what was used; if not, they went with something that felt right, only rarely more complicated than a first or second declension. The rest of the word is rendered in Latin's limited phonemes as carefully as the speaker cares to, which varies over time and according to class; for Greek words educated Romans went so far as to introduce whole new phonemes (/y/, the three aspirated stops, and also, less accurately, /z/), but nobody else got that kind of consideration, and the unwashed masses never bothered either way.
The Irish word "dún" means a "fort" and has meant that since the Early Irish period as far as I am aware, also Modern Welsh "dinas" means a city (Old Welsh "din"), both the Irish and Welsh words appear in placenames usually suggesting a fort, or fortified hill or promontory. I wonder if the Gaulish word "dunom" generally meant 'fort' too during the Roman period, cf. Lugudunom "The fort of Lug". So although probably originally meaning "a hill", like English "a down", I don't know if there is evidence that the Celtic word was borrowed into Germanic or if the Germanic words are Indoeuropean cognates. Anyway, I don't think the Germanic words mean 'fort' and just mean 'hill' so I would argue that the Romans wouldn't have borrowed it to mean "dune" from the Gauls, rather if they had borrowed the Gaulish word they would have borrowed it for "hillfort". And if it acquired the meaning "sanddune" in the Germanic languages then the Romans would have borrowed it from them. In medieval Latin I assume the Romance form "duna" (Sp. It.) could have been used as it could be easily treated as a feminine noun like 'rosa'. All just speculation on my part though!