Just noticed this thread and of course since it was relating to exploits I had to click on it and reply no matter what the question is... So here I am, the ghost of MT has returned from out of his rock to reply to you!
I'm afraid that there is no definitive answer to this question... Every email client may work differently and implement the features differetly, there is no way to class them all as one with a specific level of how vulnerable they are. In fact, you cannot just take one of the e-mail clients and say: this product is exploitable on a level of likely to be exploited.
Let me ask you a question to help yourself make an estimation guess on how secure they are:
how often do you see attachments escape the e-mail (without further user-interaction) and infect the system due to exploitation of the e-mail client?
Chances are you've never seen that actually happen to yourself, since it'd be extremely rare and tricky to actually carry out successfully. Normally the e-mail clients don't touch the attachments properly in a way to be executed - they are either stored on the cloud server and downloaded after user-interaction to the host system or they are auto-downloaded to the host system but not executed until you manually run the attachment...