New Herbert lab Nature paper reinforces mitochondrial replacement Achilles heel

Today a new Nature paper from Dr. Mary Herbert’s group in the UK has found a key problem with mitochondrial replacement therapy that fits with data from others.

Mitochondrial replacement data

The goal of preventing mitochondrial disease using various kinds of genome transfer technologies is a noble one, but mitochondrial replacement therapy has faced a number of technological challenges including perhaps most prominently the carryover and subsequent amplification of diseased mitochondria.

This mitochondrial replacement carryover issue was clearly defined in an excellent recent Cell Stem Cell paper from Dieter Egli’s lab, which demonstrated some cases of pronounced carryover and/or amplification of certain mitochondrial types following nuclear transfer. For more on my perspectives on the Egli paper see here.

In the data from the Herbert group paper, the same kind of problem is evident despite the use of a somewhat different adaptation of this technology in which pronuclei are transferred instead of nuclei or spindles from the mom-to-be. In the new paper, Hyslop, et al., the data point toward mitochondrial carryover as a significant problem still to overcome. Some scientists and politicians in the UK successfully pushed for governmental approval there to proceed with mitochondrial replacement therapy in humans last year, but I argued that this technology was not ready for prime time yet due to a number of gaps in our knowledge of both mitochondria themselves in germ cells and also what happens in actual human embryos post-genome transfer.

Now Hyslop and colleagues found that by trying out different methodological iterations they could somewhat lower the rate of mitochondrial carryover using pronuclear transfer, but it still persisted in a number of cases. It is encouraging that by changing donor-versus-mom-to-be cell freezing, altering the use of sucrose, etc. that they could bring down this carryover, but the fact it was still there is a serious concern.

Of the five hES cell lines they made with embryos after pronuclear transfer, one (#36) exhibited a pronounced rate of carryover that got worse with passaging, one (#47) exhibited a low but consistently detectable rate of carryover perhaps modestly increased with passaging, two were more encouraging (#31 and #45) and one (#55) exhibited lower rates but wasn’t cultured more than 3 passages (see Figure 4 above).

In my view, together this new Herbert group paper along with Egli’s relatively recent paper (note that the Egli paper is not cited at all in the new Herbert group paper, perhaps due to timing issues of publication although this is very unfortunate) clearly indicate that the field is not ready to use this technology to create actual people. It would be reckless to do so now without getting more data first. To their credit, Herbert’s group acknowledges the challenges in discussing their data in what is a very important paper.

One of the other challenges with genome transfer as an approach to mitochondrial disease prevention is that in a sense you have to somewhat “fly blind” as you are making a new human being this way. Unlike in these in vitro papers where readout assays can be done on whole embryos or on the ES cell lines made from them, in a reproductive context you don’t have that luxury because if you analyze a full embryo or use it to make ES cell lines, it is then gone and you have no embryo with which to proceed to try to make a mitochondrial disease-free person. You could do PGD on developing embryos that have undergone genome transfer, but even then perhaps you are just sampling six out of a 100 or so cells per embryo and could miss issues present in the majority of cells.

The overall take-home message here from this Herbert group manuscript and the one from Egli’s group is that mitochondrial replacement therapy is not ready for use in humans and substantial additional data over the next few years is needed first before even considering proceeding.

I believe it also concretely shows that the legislative approval of this technology for use in humans in the UK last year, based in part on vigorous claims from proponents in the UK that there were plenty of data already, was more political than scientific. And, yes, I’ll probably get in hot water for saying this.