The idea of an upgradeable gene pack may sound like science fiction, but it is not as far-fetched as it initially seems. In his thought-provoking book, Redesigning Humans, the biotech entrepreneur Gregory Stock describes a scenario – based on an extension of present-day technologies – that would introduce precisely this kind of flexibility into germline genetic interventions.  Stock’s argument hinges on the technology of artificial chromosomes, which were first developed in the late 1980s and early 1990s, using the genomes of bacteria and yeast.  An artificial chromosome is in many ways nothing but a pared-down version of a natural chromosome: it has telomeres on the ends, a centromere at the mid-point, and sequences of DNA base pairs along its main structure. The difference, of course, is that scientists have chosen which specific DNA base pairs to load onto the chromosome: they have assembled it, piece by piece, using a variety of recombinant technologies.
In 1997 scientists announced the creation of the first human artificial chromosome – a relatively small construct comprising about six million DNA base pairs. (Most human chromosomes are ten to forty times larger.) The breakthrough was significant, the study’s authors noted, because it opened the door for the development of new insertional vectors “capable of introducing and stably maintaining therapeutic genes in human cells.”  By 2011 scientists had coaxed artificial chromosomes to function successfully within human embryonic stem cells.  This remains very much an experimental technology, to be sure, but in Stock’s view it holds singular promise for the long haul:
Adding a new chromosome pair (numbers 47 and 48) to our genome would open up new possibilities for human genetic manipulation. The advantages of putting a new genetic module on a well-characterized artificial chromosome instead of trying to modify the genes on one of our present 46 chromosomes are immense. Not only could geneticists add much larger amounts of genetic material, which would mean far better gene regulation, they could more easily test to ensure that the genes were placed properly and functioning correctly.
Because an artificial chromosome provides a reproducible platform for adding genetic material to cells, it promises to transform gene therapy from the hit-and-miss methods of today into the predictable, reliable procedure that human germline manipulation will demand. 
Stock acknowledges the many hurdles that would have to be surmounted before such a construct could become practical. Scientists would need to make sure the added genes on chromosomes 47 and 48 did not interfere with the functioning of existing genes. (Down syndrome, for example, is caused by the presence in the genome of an extra copy of chromosome 21 – which suggests that even adding too much of existing forms of genetic material can yield significant developmental abnormalities.) Extensive animal trials would be required before even the most rudimentary extra chromosome could ever be inserted into a human.
Nevertheless, Stock argues, the key advantage of such a technology would lie in the flexibility it would confer. An artificial chromosome could be loaded with chemical switches that allowed specific genes to be turned on or off at will, simply by taking a pill containing the right chemical trigger for activation or deactivation. In addition, the entire chromosome could itself be designed in such a way as to allow it to be turned off selectively in a person’s sex cells – thereby ensuring that the construct would not be passed on to the next generation.  In other words, by coupling the technology of artificial chromosomes with the regulatory controls available through chemical interventions (ie. taking a pill with a trigger chemical), one would get the best of both worlds: genetic alterations that affected all cells in a person’s body, but that could still be tinkered with or completely shut down at any point in the person’s lifetime. These would be germline modifications, introduced at or near the moment of conception, but they would not be unchangeable or irreversible. They would allow each generation to introduce into its offspring the most up-to-date genetic modules available at the time – or to opt out of such interventions altogether, thereby reverting back to their unmodified inheritance if they so desired. Thus, these kinds of technologies would bring a crucial element of ongoing choice and flexibility into human genetic engineering.
It is worth emphasizing that this flexibility, as envisioned by Stock, would only apply at the moment of transition from one generation to the next, when parents design the germline of their offspring. It would not apply to individual adult humans: each of us would still be stuck for our entire lifetime with the gene pack engineered into us at conception. The only possible element of flexibility for individuals would perhaps be a feature that allowed people to turn off their artificial chromosomes entirely, by means of a chemical trigger taken later in their lifetimes.
 John Harrington, et al., “Formation of de novo centromeres and construction of first-generation human artificial microchromosomes,” Nature Genetics 15 (April 1997), 345-55. The quotation is from page 353. See also Joydeep Basu and Huntington Willard, “Human Artificial Chromosomes: Potential Applications and Clinical Considerations,” Pediatric Clinics of North America 53 (2006), 843-53;.
 Stock notes that scientists are already using the antibiotic tetracycline today as a chemical trigger for turning genes on and off in laboratory experiments. The prominent geneticist Mario Capecchi, moreover, has explored a specific chemical pathway through which an artificial chromosome could be deactivated selectively in human sex cells. Capecchi’s method would be to bracket the artificial genetic sequences with special segments of code, known as loxP sites, that act a bit like parentheses, marking all the DNA between them for deletion. When a specific chemical trigger (an enzyme known as CRE) is administered, this would activate the loxP sites and cause the artificial chromosome to be snipped out and deleted. In Capecchi’s words:
At the same time as new information is introduced into the germline, the loxP sites and CRE recombinase gene needed to reverse the change would also be introduced. This would, at the patient’s discretion, allow subsequent deletion of all information introduced into his or her germline. … The CRE recombinase could be accompanied by control elements to allow it to be activated in response to a drug taken by the patient, which would result in deletion of essentially all of the added information from his or her germ cells. Thus, the added information would not be transmitted to subsequent offspring.
Capecchi has subsequently confirmed, in experiments with mice, that this chemically-triggered deletion mechanism works reliably. Mario Capecchi, “Human Germline Gene Therapy: How and Why,” in Gregory Stock and John Campbell, eds., Engineering the Human Germline: An Exploration of the Science and Ethics of Altering the Genes We Pass to Our Children (Oxford, 2000), 38-39.