Author response to Commentary on "Long-term safety and efficacy of deoxycytidine/deoxythymidine in treatment of POLG-related disorders"

We appreciated the commentary by Hirano et al. [1] on our recently published paper, “Long-term safety and efficacy of deoxycytidine/deoxythymidine in treatment of POLG-related disorders” [2]. This established group of clinicians and researchers made some important points, which we generally felt helped put our findings in proper context. We also feel their comments highlight some broader issues regarding the current state of rare disease therapeutic research and the differences between industry-sponsored studies and studies from academic centers funded by grants from non-profit agencies.

Regarding the evidence for efficacy of pyrimidine deoxynucleosides for individuals with POLG-related disorders, we agree that any data from an open-label study should be interpreted with caution. It is true that the Newcastle Mitochondrial Disease Scale has some subjectivity, but this is the case for other disease-specific severity scales that have been recently used in industry-sponsored trials. A good example is the Angelman syndrome Clinical Global Impression Scale [3] which has been adapted for use in several recent phase 2 open-label clinical trials. We also emphasize that our results include improvements in more objective outcomes, such as serum growth differentiation factor 15. As well, we have previously presented data demonstrating an improvement in delta power on electroencephalography in our broader cohort of patients with mitochondrial DNA (mtDNA) depletion disorders (manuscript in preparation) [4].

From a mechanistic perspective, there is more evidence than was mentioned in the commentary for pyrimidine deoxynucleoside safety and preserving mitochondrial integrity in POLG and other non-TK2 mtDNA depletion disorders. In the fibroblast study by Dombi et al. cited in our paper, the effects of many different deoxynucleoside combinations were compared [5]. The authors reported “Supplementation of deoxy pyrimidines alone or in combination (C, CT or T) did not reduce mitochondrial membrane potential or cell numbers and were hence harmless to the cells.” This observation was in contrast to evidence for potential toxicity of various nucleoside combinations including deoxyguanosine, leading Dombi et al. to propose that the specific combination of deoxycytidine and deoxythymidine “could be explored further for treatment of POLG patients.” In the text of the paper, they also reported that “CT significantly increased mtDNA content in all patients and controls (log transformed summed area of mtDNA nucleoids compared with baseline, POLG2 p < 0.01, the rest all p < 0.001 with LMM), as did T or C alone in DGUOK cells (p < 0.01 with LMM).” There is also a study by Kristiansen et al. that evaluated the impact of deoxyribonucleosides on a POLG induced pluripotent stem cell-derived neural stem cell model that showed reversal of mtDNA depletion, though they only evaluated the effect of all four deoxyribonucleosides together [6].

In general, we agree that clinical trials, particularly those for rare diseases, should be based on a plausible mechanism, ideally supported by robust pre-clinical data. However, it is also worth noting that medicine has a long history of developing treatments for one disease that end up being as effective or more effective for other conditions. As a relatively recent example of the pitfalls of hypotheses based on in vitro data, there was considerable excitement over the use of memantine, a medication approved for dementia for which the primary mechanism of action was thought to be NMDA receptor antagonism, for patients with epileptic encephalopathy (EE) due to gain-of-function pathogenic variants in genes encoding NMDA receptor subunits such as GRIN2A [7]. However, we conducted a randomized placebo-controlled cross-over trial that showed that memantine could also be effective for patients with EE due to a wide range of etiologies, even those with loss-of-function variants in NMDA receptor subunit genes, suggesting that the mechanism of action is much more complex [8].

Experiences such as this emphasize that, although some degree of skepticism is reasonable when apparently surprising clinical responses are observed, if the data are compelling the next step should be to investigate other possible mechanisms by which the intervention might elicit clinical benefit. Our paper [2] mentions alternate possible therapeutic mechanisms, such as reversal of premature telomere shortening, and we are currently planning experiments to better understand underlying molecular mechanisms that might lead to therapeutic benefit of pyrimidine deoxynucleosides in POLG-related disorders. It is also worth noting that reversal of mtDNA depletion has not actually been demonstrated in humans for TK2 deficiency; rather, this is a presumed mechanism based on data from in vitro and mouse models [9].

Finally, our experiences with this study have highlighted the challenges of conducting rare disease clinical trials without the greater financial support of a pharmaceutical company. When dependent on the less certain funding support of non-profit organizations, it is much more challenging to make long-term plans for clinical trials. At the point when we had positive clinical data from only a few patients, we considered a phase 3 double-blind randomized controlled cross-over trial which would have generated a higher level of evidence. However, we were unable to secure financial backing from government granting agencies. At this point, such a trial may now be ethically unfeasible, at least for early-onset POLG-related disorders, as a significant proportion of patients randomized to placebo would likely die during the study.

Author contributions

KAM drafted the initial manuscript. PJW and DB reviewed and revised the manuscript.

Declarations of interest

None.