Choosing interventions

Our primary interest when determining which interventions to include in RMR studies is, of course, the strength of pre-existing evidence pointing to damage reversal. For RMR1 our criteria were extremely stringent: that treatments have already demonstrated lifespan extension individually, and that they did so when begun in mid-late life in wild-type mice. This accomplishes a couple of things. Firstly, any intervention which is capable of extending lifespan in this way must, by definition, be addressing/alleviating all types of damage which cause significant morbidity, through a combination of direct and knock-on effects. The task, then, is to determine if we can extend this further by combining interventions which do this, but through different mechanisms – which introduces a third criterion, namely that the interventions should be highly divergent in their presumed direct targets.

For RMR2, our criteria are less stringent, though still focused on damage-reversing therapies. We remain interested in interventions which meet the previous criteria, but now we also consider treatments which may not have clearly established maximum lifespan benefits, but which significantly improve healthspan and/or mean lifespan. In practical terms this differs from RMR1 in that therapies only affecting healthspan are likely to primarily, and robustly, address a single or few drivers of pathology, but don’t appear to have enough systemic knock-on effects to cover all bases and thus increase maximum lifespan. A combination of such treatments, however, might achieve full-spectrum damage repair – perhaps even more so than treatments affecting every system to some more minor degree.

Duration of benefit

The approximate duration of benefit from a single administration of a given intervention is also important to consider. This is especially relevant when selecting more invasive interventions, such as those requiring intravenous or intraperitoneal injection of material. Not only is that process physically and psychologically stressful for animals, it also comes with the risk of complications, particularly when anesthesia is required. Furthermore, invasive treatments which must be given frequently greatly increase the required reagent amounts and technician time required, in addition to being less likely to be therapeutically practical in humans. Therapies which are effective when given infrequently or intermittently are thus greatly preferred when possible.

Translatability

While our combination studies are, in a sense, proof-of-concept for combined intervention effects, we prefer to avoid therapies which are unlikely to have any path to the clinic in the foreseeable future, for example heterochronic parabiosis. This does not preclude the inclusion of intervention variants, however, based on the theme of a promising intervention. We will discuss this further below in the context of saline albumin.

Technical feasibility and invasiveness

There are studies which have yielded impressive results, but which cannot be practically replicated (within a budget that we can consider realistic) at the scale necessary for statistical significance in a multi-intervention animal study. For example: therapeutic plasma exchange or plasma dilution, as previously used, require animals to have surgically implanted cannulae for administration, and typically require a large number of treatments because of the limiting blood volumes which can be exchanged at once. Studies using ‘young plasma’ source plasma directly from sacrificed donor mice, and even those giving only saline-albumin require sacrificed donor mice to supply the red cell component. Any cell or fluid preparation harvested directly from donors is typically isolated and processed immediately before administration to recipients. Logistically, this can be quite challenging even on a small scale, let alone a study of this size. In the context of HSC transplant from young donors in RMR1, we navigated this challenge by staggering the treatment days and groups from study start, as the study mice were only planned to receive the treatment once. Still, to collect bone marrow for a one-time treatment in 1000 mice, an entire team of technicians had to be brought in from another lab site for 3 separate 5-day spans. Therefore any treatments which we aim to include in RMR2 are considered on the basis of commercial material availability or ease of manufacturing, in conjunction with biological implications of bulk manufacture, for example, stem cell behavior. Treatments which can be administered in the animal chow have the benefit of permitting continuous, non-invasive dosing and reducing animal handling and stress. For this and other reasons, most studies investigating intervention combinations (or even single interventions), including studies conducted by the NIA’s Intervention’s Testing Program (ITP), are restricted to testing orally-available compounds. Because orally available therapies already receive sufficient attention in lifespan studies, we are largely interested in those treatments which are more difficult to administer, and thus more complex. Unfortunately, any treatment materials which are administered in a way other than orally or topically are considerably more expensive than those which can be consumed, as products must be sterile and GLP manufacturing or similar grade is often necessary or required. But they include essentially all gene and cell therapies, so omitting them amounts to asking for failure. Additionally, for any therapies which are administered in chow, it must either be acceptable that the amount of a molecule an animal consumes is unknown, or that animals can be singly-housed with rationed feeding. While single animal housing would enable precise dosing of chow-administered treatments, the negative health outcomes linked to social isolation in rodents are undisputed and fail to justify single housing solely for control of food consumption. Investigation of dietary interventions such as caloric restriction can still be, and often are, thus conducted in group-housing conditions through the use of reduced-calorie food rather than controlled consumption volume, which is a decidedly approximate proxy for the real thing.

Underrepresentation in research

As in the above example of orally available molecules being heavily studied by the ITP, we have chosen to focus on therapies with strong evidence, but which are not already extensively covered by the field. This primarily concerns combinations of small molecules such as rapamycin, metformin, resveratrol, NMN, dasatinib + quercetin, fisetin, acarbose, captopril, and similar. In RMR1, our exception was to include a single well-studied intervention, rapamycin, which consistently extends rodent lifespan and thus also serves as a control, allowing us to compare outcomes from RMR with other lifespan extension reports and better gauge any synergies which may emerge from combination therapies.

Financial feasibility

Unfortunately, some very promising interventions would just not be financially justifiable for a study of this scale. For example, in our preparations for RMR1 we received quote approximations for an exciting therapy which would have cost more than $1M to manufacture for 500 mice, just for a single dose. We also carefully consider options for what form an intervention might take, how frequently it would need to be given, how much of a product is needed to achieve therapeutic benefit, and whether the act of repeated dosing is likely to harm the aged animals. Any administration requiring specialized technical skills incurs a labor cost, and thus those which would need to be given weekly or more frequently would significantly increase overall study costs, in addition to material costs. We are cognizant of how a material source can impact the bottom line for cost of a therapy, weighing costs of manufacturing options against potential consequences, such as diminished efficacy or bandwidth restrictions. Additionally, therapies which would require single-animal housing would increase the cost of basics like the number of cages needed, the amount of nesting material, the hours required for cleaning, etc. in the same way that very frequent treatments increase the cost of materials and technician labor.

Last updated