HDAP2

HDAP2 (high-density aromatic peptide 2) comes from Alexander Birk's lab at York College, City University of New York [1]. Birk was first author of the 2013 study showing that SS-31, now the drug elamipretide, acts by binding cardiolipin [4], and HDAP2 goes after the same lipid by a different route. Its core is two phenylalanines side by side. Earlier peptides built around a Phe-Phe core damaged mitochondria and killed cells, so HDAP2 brackets the pair with a D-arginine on each side, which was meant to make it water-soluble and non-toxic [1]. The biotin on its N-terminus was attached so the peptide could be traced in tissue [2]. The patent filing also covers related peptides, among them HDAP4 (D-Arg-Phe-Phe-Phe-D-Arg) and HDAP6 (D-Arg-Trp-Trp-D-Lys), and claims their use in the eye for glaucoma and optic nerve injury [5].

Cells. In kidney cells (MDBK) deprived of serum for 5 days, a stress that depends on mitochondrial failure, HDAP2 kept the cell layer intact and raised viability up to 1.8-fold, with half its maximal effect at about 100 nM. At 1 µM it kept mitochondrial membrane potential about 2.5-fold higher than in untreated cells and cut an oxidative-stress signal by about 80%. Biotin-tagged HDAP2 entered the cells within an hour, about six times faster than the biotin compound biocytin, and gathered on mitochondria [1].

Optic nerve crush. Mice were given 3 mg/kg a day by injection into the abdomen, starting right after one optic nerve was crushed, or saline. After 14 days, ganglion-cell density was about 630 cells/mm² in treated retinas against 446 in the saline group, 41% higher, with similar gains in the center, middle and edge of the retina. Both groups had still lost most of their ganglion cells: uninjured retinas averaged about 3,270 cells/mm². In the crushed nerves, treated axons held 2.8 times as many mitochondria as untreated ones, about 62% of the uninjured level, but the mitochondria that remained looked the same in both groups, so the authors conclude that HDAP2 limits the loss of mitochondria rather than repairing damaged ones. Axon counts were higher with treatment but not significantly. In mice without injury, HDAP2 left ganglion-cell numbers unchanged [2].

Glaucoma model. DBA/2J mice develop raised eye pressure and lose ganglion cells as they age. Mice given 3 mg/kg every other day from 4 months of age until they were examined at 10 to 12 months (32 retinas) were compared with untreated mice from the same colony (26 retinas). Eye-pressure exposure was the same in both groups, so the peptide does not lower pressure. Treated retinas kept about 38,800 ganglion cells against 26,400, preserving about half of the cells that untreated eyes lost (young controls had 51,700). The difference grew with pressure: none at low exposure, 58% more cells at moderate exposure and 2.8 times as many at high exposure. Severe loss, below 15,000 cells, occurred in 16% of treated eyes (5 of 32) against 46% of untreated eyes (12 of 26), and treated eyes tolerated 29 mm Hg more cumulative pressure before reaching it. Treated optic nerves had about 16,000 axons against 7,500 in untreated nerves, and looked like those of young controls. Even so, ganglion cells in treated eyes still declined as pressure rose, and the slope of that decline did not differ significantly from untreated eyes [3].

What has not been done. The untreated glaucoma mice were not given saline injections, which the authors explain by the strain's sensitivity to handling, and neither mouse study tested vision itself [2][3]. All of the work comes from one group, whose two lead authors are inventors on the patent; they report no licensing or commercial ties [2]. There is no human study.

Cardiolipin is a lipid found almost only in the inner mitochondrial membrane. Its shape lets the membrane bend into the folds (cristae) where ATP is made, and it binds and supports several respiratory-chain proteins. Under stress, such as raised calcium or oxidized cytochrome c, patches of cardiolipin leave the flat bilayer for inverted hexagonal and micelle shapes. In those non-bilayer phases it promotes mitochondrial fission and mitophagy, forms a complex with cytochrome c that acts as a peroxidase, oxidizing the membrane and releasing cytochrome c to trigger apoptosis, and lets protons leak, which lowers the membrane potential and ATP output. HDAP2 was designed to keep cardiolipin in the bilayer [1].

Binding. In liposome experiments HDAP2 displaced a cardiolipin-specific fluorescent probe from liposomes containing cardiolipin, at about one peptide per two cardiolipin molecules, and did nothing to liposomes made of phosphatidylcholine alone. Adding cardiolipin made the liposome membrane more polar, as read by Nile Red, and HDAP2 reversed that shift. With pure cardiolipin, which forms inverted micelles in water, HDAP2 produced vesicles 1–3 µm across. By analogy with the SS peptides, the authors propose that the phenylalanines sink into cardiolipin's fatty chains while the arginines hold its phosphate head groups; there is no NMR or cryo-EM structure yet [1].

How it differs from SS-31. SS-31 binds cardiolipin one-to-one, and its dimethyltyrosine scavenges free radicals. HDAP2 binds about two cardiolipins and has no scavenging group, so its drop in oxidative stress is thought to be indirect: keeping cardiolipin in the bilayer should stop cytochrome c turning into a peroxidase. The authors say the cytochrome c part of this is still indirect evidence. Because HDAP2 carries a net positive charge, they expect it to bind cardiolipin from outside the mitochondrial matrix [1].

In the retina. After injection, 77% of the peptide's signal in the inner retina overlapped a mitochondrial marker, and much of it was in Müller glia as well as ganglion cells [2]. Part of the protection may therefore come through glia, and cardiolipin stabilization itself has not yet been shown in animals [3]. In crushed optic nerves it preserved the number of mitochondria rather than their structure, which the authors interpret as preventing mitochondria from being removed after injury [2].

Every published study used biotin-D-Arg-Phe-Phe-D-Arg-amide, synthesized commercially at over 96% purity, with the biotin added as a detection tag [2][3]. The developers' patent filing reports that the tag matters: at 1 µM the peptide without biotin did not protect serum-starved kidney cells. In mice with itching caused by a mitochondrial poison applied to the skin, biotin-HDAP2 put on the skin reduced scratching while the biotin-free peptide increased it 2.5-fold (3 mice per group). The filing puts this down to the structure of the biotinylated peptide without explaining it [5]. These results have not been published in a peer-reviewed paper.

HDAP2 has not been given to people, so there is no evidence about its safety in humans either way. In cultured kidney cells it caused no toxicity up to 100 µM, about 1,000 times the concentration that gave half its protective effect [1].

In mice the toxicity is clear and dose-dependent. In a 14-day study of daily injections into the abdomen, 300 mg/kg caused seizures and death within minutes, 100 mg/kg caused sedation, seizures and death within 30 minutes, and every mouse on 50 mg/kg was dead by day 9. All mice on 3 or 10 mg/kg survived with normal weight and behavior. The LD50 was 55 mg/kg and the maximum tolerated dose 30 mg/kg [2], so the protective dose of 3 mg/kg sits about 18-fold below the LD50 and 10-fold below the maximum tolerated dose.

At 3 mg/kg it did not change ganglion-cell numbers or retinal layout in uninjured mice, and the authors report the same dose was tolerated through 8 months of injections in glaucoma-prone mice [2]. How the body handles the peptide has not been studied [1].