Targeted cell killing technologies are one of the most important developments to emerge from the cancer research community. Beyond the immediate target of cancer cells there are in fact a whole range of specific types of cell that we'd like to periodically eliminate from the body, safely, and with minimal damage to surrounding tissue. Senescent cells, for example, contribute directly to the aging process as they accumulate with the passing of years. Also, the immune system fails in part because it has a limit on the number of cells it can support, and too many of those cells become uselessly specialized to detect and combat mild herpesviruses like CMV that cannot be cleared from the body by its own natural processes. Getting rid of those cells would free up space that is very much needed.
All of these cell types have their characteristic differences, and given a reliable way to take advantage of those differences then some form of targeted destruction can be unleashed to improve health and reverse this aspect of aging. Two of the more popular approaches to targeted cell destruction in the cancer research community involve nanoparticles such as dendrimers and viruses. The former is a bottom-up approach to building a tool: the construction of comparatively simple, minimalistic assemblies that are designed to link together and deliver a collection of specific designer molecules - perhaps a sensor to match to a type of cell, something to cause the cell to ingest the particle and its payload, and something that will sabotage the cell. Viruses on the other hand are much more complex entities, and their use embodies more of a found tool approach: some are useful because they have a preference for cancer cells over normal cells. Others can be tailored to act that way, but you work with what you can find in the wild or breed in the lab.
As researchers build better nanoparticle-protein assemblies and become more adept at manipulating viruses to exhibit specific desired behaviors, the line between the two will eventually blur. Viruses are about as close as you can get to chemistry while still being something that is generally accepted as being biology. There's a way to go yet - no-one is producing self-replicating nanoparticles for medical use that I'm aware of - but it will happen.
Here's an example of ongoing work at the virus end of the community, showing how researchers are becoming more readily able to alter the mechanisms of viral activity to achieve the desired result, in this case improving its ability to target cancer cells and only cancer cells:
Newcastle disease virus kills chickens, but does not harm humans. It is an oncolytic virus that hones in on tumors, and has shown promising results in a number of human clinical trials for various forms of cancer. However, successful treatments have required multiple injections of large quantities of virus, because in such trials the virus probably failed to reach solid tumors in sufficient quantities, and spread poorly within the tumors.
The researchers addressed this problem by modifying the virus's fusion protein. Fusion protein fuses the virus envelope to the cell membrane, enabling the virus to enter the host cell. These proteins are activated by being cleaved by any of a number of different cellular proteases. They modified the fusion protein in their construct such that it can be cleaved only by prostate specific antigen (which is a protease). That minimizes off-target losses, because these "retargeted" viruses interact only with prostate cancer cells, thus reducing the amount of virus needed for treatment.
Retargeted Newcastle disease virus has major potential advantages over other cancer therapies. [Its specificity for prostate cancer cells means it would not attack normal cells, thereby avoiding the various unpleasant side effects of conventional chemotherapies. In previous clinical trials, even with extremely large doses of naturally occurring strains, "only mild flu-like symptoms were seen in cancer patients."