Just an opinion here and not canonical advice, but you're headed for a world or hurt if you're trying to extract exact text values from the DOM using the inner/outer HTML/TEXT properties via Javascript. Different browsers are going to return slightly different values, based on how the browser "sees" the internal document.
If you can, I'd change the HTML you're rendering to include a hidden input, something like
<table>
<tr>
<td id="MyCell">Hello World<input id="MyCell_VALUE" type="hidden" value="Hello World" /></td>
</tr>
</table>
And then grab your value in javascript something like
document.getElementById(cell2.Element.id+'_VALUE').value
The input tags were designed to hold values, and you'll be less likely to run into fidelity issues.
Also, it sounds like you're using a .NET control of some kind. It might be worth looking through the documentation (ha) or asking a slightly different question to see if the control offers an official client-side API of some kind.