You don't really appreciate just how awful scientific reporting is until it's about a field you've worked in. (But it's also
really hard to break down and generalize, I've learned.) Let me apologize in advance.
Anyway, this is
not a new type of T cell. T cells have a receptor called a 'T cell receptor', or TCR. Usually, these TCRs recognize broken down fragments of proteins, which triggers an immune response. One such response is killing whatever cell that activated it. The great thing about TCRs is that they are extremely diverse, giving your immune system the ability to handle almost any threat, even if it's something it has never seen before. The benefit of transplanting in T cells that have been engineered to target a disease is that T cells can persist in a "memory" form and then grow up again if that same threat appears again, giving you protection over time. (Think of how vaccines "prime" your immune system so that your body is able to immediately respond to the real threat later.)
Usually, immunotherapy research starts with finding an attractive target. For example, maybe some cancers express a lot more of a certain protein compared to healthy cells, or they have the same protein in a mutated form. However, they seem to have found a TCR that recognizes some
unknown target that was present on the cancer cells they used (across different cancer types) but left normal cells alone. TCRs are very specific. But there is also the issue of a TCR having the potential to be cross-reactive with normal cells
that you have yet to test. (This could be tissue types you have not accounted for, or people with different versions of proteins due to having different versions of genes. Immunotherapies tend to be developed with majority populations in mind.)
What could be interesting about the TCRs they have found is that they do not depend on the "classical" recognition mechanism. Something I left out is that a TCR generally cannot recognize a protein that has not first been "processed" into its constituent parts and
presented by the target cell (on its surface) to the T cell. This adds another layer of specificity that means that certain TCRs will only work in certain people. This specificity (if you read about how "HLA" works) is the basis of transplant rejection and graft versus host disease. What the authors hope is a class of T cell that won't be limited to certain populations. (Pan-cancer TCRs are already a goal, so that's not anything particularly "new".)
This is not remotely close to testing in people. First, they don't even know what their TCR is truly targeting. Second, good data in mouse or in cell cultures does not automatically equal a good therapy in people. We have had very promising looking TCRs that were bunk in practice. We have had TCRs that more or less cured two patients, but had no response in the rest without any clear reason why. Cancers have many different strategies to subvert or avoid immune surveillance; you can wipe out the majority of cells with a protein you're targeting, and then a small percentage without it (that therefore survived) can grow out that will be resistant to that same therapy. Liquid tumors (leukemias) are easy to deal with; solid tumors require T cells to infiltrate into an environment where they can't be effective. That's a huge issue that remains to be figured out. There is a lot of safety testing that needs to be done since the
consequences can be fatal. (CAR-T cell therapy, especially, has a huge concern with neurotoxicity.) The real truth is that we don't really know what to expect going into a lot of these trials, though due diligence is critical.
(I genuinely tried to keep this as simple, concise, and straight-forward as possible, but neither cancer nor the immune system are any of those things. Oops.)