1) Competetive elements: The fact that a player has completed a dungeon a few times and gotten all the loot from it doesn't have to mean that it is pointless to go back, you can add rankings for fast clears(with standardized gear and levels this can also keep content intended for low level players interesting for high level players) and hand out rewards to the fastest x or x% players each week, similar things can be done for both solo and group content.

2) Social elements, Large scale projects that guilds or the server community as a whole can participate in. (Look at things like the community collections in Ultima Online for example), These things can interact with PvP elements aswell by giving advantages to the faction or guilds participating.

3) Community organized adventures, (Diffcult to implement) allowing a player to set up their own instanced adventure in a zone or dungeon for friends/guildmates or anyone who happens to be interested, placing enemies, configuring objectives and even controlling npcs can create some interesting possibilities.

I pretty much agree with wyrmslayer - all games get boring unless you are basically providing new game segments. But I'll make some overall comments on what I think makes a good PvE MMO.

First of all I strongly prefer games that have npcs which deliver story content to orient and immerse the player, and form the foundation of the social fabric that players interact with each other and roleplay within. These NPCs can be multifunctional if also used as part of a dating sim relationship building system, a factipn points grinding system, and to provide various in-game services. Its much better to have fewer than 50 NPCs who each can interact with the player in multiple ways than it is to have hundreds who only have one thing to say.

Second, player crafted housing and gardens are awesome. There's no reason they have to exist in the main game world if this is a technical hurdle; Wizard 101 for example has each house as a little island in space and it works great. Along the same lines customizations like mounts, pets, clothing, tattoos, hairstyles and even anthropomorphization are some of the most long-term motivating rewards any MMO can use to keep players playing for months. Not to mention these kind of things motivate the player to explore for gatherable resources and recipes as well as fight to drop-hunt crafting mats.

Third, combat - WoW-style combat is inherently boring because the monsters have to be pretty homogenous to be equally killble by all classes, plus the player has to act fast (limiting thinking) and there are quite limited tactical options wihin each class. I much prefer combat like that in Dofus, a turn-based tactical combat paired with a system where players use points to purchase their choice among an array of spells. it would be even better if there were no classes so all spells were available to one character, and the player could easily switch between different builds which were effective against different monsters. For example they could equip an invisibility spell to avoid being poisoned by a cobra, but a bat wouldn't be fooled an the player should instead equip a speed bonus or a paralyzing projectile to catch the speedy agile bats.

Fourth, mini-games and a forum, build them right into the MMO. NeoPets and Gaia Online are two examples of how an online game can keep players interested for years without even having a world the players can walk around and kill monsters in. Good minigames, for example classic solitaire, have tons of replayability. (Gaia does now have a more standard monster-killing world but that's a fairly recent add-on).