21 Jun 2024

Mutations that douse expression of the progranulin protein underlie nearly a third of familial frontotemporal dementia cases. Could scientists simply replace what is missing? Multiple progranulin replacement therapies are vying to do just that. Two new studies showcase different delivery strategies.

One uses a so-called protein transport vehicle (PTV)—essentially an antibody trained against the transferrin receptor—to whisk an attached progranulin payload across the blood-brain barrier. Scientists led by Christian Haass, Anja Capell, and Dominik Paquet of Ludwig Maximilians University in Munich, along with Denali’s Gilbert Di Paolo, expressed the PTV-PGRN combination from an AAV8 virus. It targets the mouse liver, from where it pumped a continuous supply of the fusion protein into the blood stream after a single injection. In the blood the chimera floated into the cerebrovasculature, where it crossed the blood-brain barrier. Importantly, it reached every nook and cranny of the brain parenchyma. This corrected a slew of phenotypes, including lysosomal dysfunction, TDP-43 aggregation, gliosis, and motor deficits. Alzforum covered the concept when Haass presented it at ADPD 2023 (Apr 2023 conference news). The full report appeared in the June 5 Science Translational Medicine. These AAV experiments were proof of concept. For clinical studies, Denali, and its partner Takeda, inject PTV-PGRN, sans virus, into the blood. Phase 1/2 trials are underway.

The paper comes on the heels of interim findings from a Phase 1/2 trial of a different approach. Scientists led by Jeff Sevigny at Eli Lilly’s New York location injected a progranulin-expressing AAV9 vector into the cisterna magna of 13 study participants with FTD-GRN. LY3884963, known as PR006 before Lilly acquired its original developer, Prevail Therapeutics, more than doubled progranulin in the CSF. Levels held steady a year after the injection. The trial was not designed to measure efficacy, but it did record some side effects. Together, the studies suggest that both approaches can deliver progranulin to the brain. Each strategy has pros and cons.

Blood versus Brain
For the PTV approach, co-first authors Marvin Reich of LMU and Matthew Simon of Denali put their therapy to the test in mice missing both GRN and TMEM106b. Unlike GRN KOs, which have lysosomal dysfunction and inflammation but no TDP-43 pathology, the double KOs develop a range of FTD problems more reflective of human FTD, including aggregates of TDP-43 and severe motor deficits.

Out for Delivery. A liver-tropic AAV has on board a gene for mouse TfR-binding antibody fragment (8D3) fused to one encoding human progranulin (top). After injection into a tail vein, the AAV infects hepatocytes, which churn out 8D3:PGRN. It crosses the blood-brain barrier (right). [Courtesy of Reich et al., Science Translational Medicine, 2024.]

The scientists injected the AAV8 into the mice intravenously. One week later, progranulin was barely detectable in the plasma. Eight weeks later, progranulin had risen in the plasma and it had spread throughout the brain, where it held steady for the 31-week study period (image below). This broad distribution is made possible by the brain’s extensive vascularization, Di Paolo told Alzforum. Essentially, wherever a blood vessel or capillary flows, the fusion protein can, and did, cross into the parenchyma.

Because AAV infections can damage the liver, and because TfR is highly expressed in erythrocytes, the scientists checked for signs of toxicity in the blood and in multiple organs. They found none, though red blood cells were slightly smaller and more numerous.

 

The progranulin replenishment markedly improved the mice’s phenotype. While untreated dKOs promptly fell off a spinning rod and failed to right themselves after being pushed over, the treatment steadied them on the rod and bolstered them against a gentle shove. Myriad cellular problems in the brain, including lysosomal dysfunction, flagging autophagy, insoluble TDP-43 aggregates, elevated neurofilament light, and neuronal death, were rescued. Fewer microglia expressed disease-associated transcripts, and tempered expression of genes indicating lysosomal dysfunction. Numbers of reactive astrocytes and activated microglia fell. Lipid balance was restored, including levels of the phospholipid bis(monoacylglycero)phosphate (BMP), which falls when progranulin is deficient.

That’s all well and good in mice, but what about in human cells? To test this, the researchers used a human cell culture model in which wild-type iPSC-derived neurons share a dish with iPSC-derived microglia that lack both GRN and TMEM106b. The rationale is that if microglia don’t produce progranulin, neurons bear the brunt of TDP-43 pathology. Indeed, these microglia spurred TDP-43 aggregation, and ultimately, death of the neurons. Treatment with PTV:PGRN quashed TDP-43 pathology, and spared the neurons.

Fenghua Hu and Tuancheng Feng of Cornell University in Ithaca, New York, called the liver-targeted AAV approach used in the DKO mice exciting. “This novel strategy rescues all the pathology associated with PGRN deficiency with no obvious deleterious side effects in the mouse model, thus holding therapeutic promise for FTLD-GRN,” they wrote.

How often would PTV-GRN need to be administered to maintain normal progranulin levels in the brain? That is under investigation, said Di Paolo. However, a dose administered every other week was shown to correct all phenotypes of the Grn KO mouse (Logan et al., 2024). Though this approach is more cumbersome than a single shot with an AAV vector, injections of the recombinant protein can be stopped at any time should side effects develop. Even so, di Paolo thinks peripheral AAVs could have promise, as well.

The scientists also emphasized that the TfR-based therapies avoid potential pitfalls that come with directly injecting viruses into the CNS. Not only is the latter more invasive, but some viruses may unevenly distribute through the brain.

Straight into the Brain
LY3884963, aka PR006, uses a more CNS-targeted approach. It comprises an AAV9 vector carrying the progranulin gene, and is injected into the cisterna magna. This vector primarily infects neurons, but has also been demonstrated to target microglia, Prevail’s Olga Uspenskaya told Alzforum.

In their Nature Medicine paper, the scientists reported preclinical findings from GRN KO mice and nonhuman primates, as well as interim results from the ongoing open-label, Phase 1/2 study. Injected into the brain ventricles of adult GRN knockout mice, the AAV9 dose-dependently boosted expression of human progranulin in the cortex and spinal cord. Multiple cellular phenotypes, including accumulation of lipofuscin, a glut of ubiquitinated proteins, and an uptick in pro-inflammatory cytokines and gliosis, were corrected. BMP in the urine rose to wild-type levels, indicating a boost in lysosomal function, and hinting that urine BMP might be a biomarker.

Before moving into human trials, the scientists administered PR006 into the cisterna magna of cynomolgus macaques to check for toxicity. In three monkeys per group, they tried a buffer control, as well as two doses—3.9×10⁹ and 3.9 x 10¹⁰ vector genomes per gram of brain—the latter more than double the lowest dose later used in the clinical trial. There were no obvious health changes initially, though six months later, necropsy revealed a potential problem in several of the animals on PR006. They had more glial cells in the dorsal root ganglia, This suggests that glia were reacting to neuronal injury, a phenomenon that has been reported for other CNS- targeted gene therapies (Hordeaux et al., 2020). Even so, none of the monkeys had clinical signs of DRG degeneration, such as instability. Transcripts encoding the human progranulin transgene dose-dependently increased in the cortices, hippocampi, and mesencephala of the macaques, while the highest dose doubled progranulin protein in the CSF.

In the clinical trial, called Proclaim, six people were randomized to receive a low dose of 2.1 x 10¹³ vector genomes. The other seven received twice that. Plans for an even higher dose were scrapped after initial biomarker findings from the lower doses looked promising. For this open-label study, primary endpoints are safety and CSF progranulin levels, while Clinical Dementia Rating plus National Alzheimer’s Disease Coordinating Center Frontotemporal Lobar Degeneration rating scale (CDR plus NACC-FTLD) and levels of neurofilament light chain (NfL) serve as secondary outcomes. Exploratory endpoints included measurement of urine BMP.

On average, participants had been diagnosed with FTD one to two years prior to enrollment, and had mild to moderate dementia at baseline. Eight had first presented with behavioral variant FTD, one with primary progressive aphasia, and four with a combination of both. Notably, five were seropositive for AAV9-specific antibodies at baseline. Uspenskaya told Alzforum that despite the presence of these antibodies, the team decided to include these participants in hopes that in the CNS, the virus might escape detection by the immune system.

Progranulin Peak. CSF progranulin concentration increased in all trial participants in response to either dose of PR006. It then dropped, but remained within the normal range reported for noncarriers, in most participants at 12 months. [Courtesy of Sevigny et al., Nature Medicine, 2024.]

Their gamble paid off. Two months after injection, CSF progranulin levels—which were around half those typically reported for healthy noncarriers at baseline—rose 2.5- to 4.5-fold in the low-dose group, and three- to sevenfold in the mid-dose group. Among nine participants who had reached the six-month data point, eight retained normal or above-normal progranulin levels, and at 12 months, three of four participants did. Plasma progranulin took a different trajectory, rising one month after injection before returning to baseline levels by two to three months. In a post hoc analysis, however, plasma progranulin hardly budged in the five people with pre-existing antibodies to the viral vector. This suggested that their antibodies made quick work of the AAV in the blood.

The urine concentration of two BMP derivatives also rose on both doses of PR006, approaching levels in healthy controls. These phospholipids peaked at six months and stayed there at 12 months. At the AD/PD meeting in Lisbon, Uspenskaya had reported a similar uptick in the CSF.

How did participants fare in response to PR006? Each experienced an adverse event. The most common was an infiltration of white blood cells into the CSF, aka pleocytosis, which was likely caused by the treatment. This occurred in six participants, and was detected two months after injection. It resolved one to four months later. One person developed hearing loss, thought to stem from this CSF inflammation. At the time of the analysis, this participant was improving, Uspenskaya said.

CSF pleocytosis likely reflects an inflammatory response to AAV9-PGRN transduction of the sensory dorsal root ganglia, which was predicted to peak at two months. In support of this, both plasma and CSF NfL spiked around the same time, suggesting ongoing inflammation and/or neuronal injury. None of the participants noticed changes in the sensations DRG neurons transmit, such as pain, touch, temperature, spatial orientation, or vibration, during the trial.

NfL levels dropped back to baseline in most participants by six months, and, as Uspenskaya reported at the AD/PD meeting in Lisbon, it remained at or below baseline in 70 percent of participants at 12 months. Given the 30 percent annual rise in NfL that is typical of people with FTD-GRN, Uspenskaya considers this a promising result. It hints at a possible neuroprotective effect that takes over after the initial inflammatory reaction has subsided, she suggested. Participants declined on the FTLD-plus-NACC during the open-label trial, though it was not designed to tease out clinical benefits.

Other adverse events that cropped up during the trial, including osteoporosis, hypercholesterolemia, hyperlipidemia, diabetes, and prediabetes, were judged unrelated to PR006.

Three participants experienced deep-vein thrombosis, a serious adverse event. All three instances were considered unrelated to the drug by Lilly, while one was deemed possibly related to treatment by the clinical site investigators. These three patients had risk factors for blood clots. They had a high body-mass index, travel-related immobility, and took the anti-inflammatory drugs sirolimus and corticosteroids. These were included in the study protocol, to mitigate immune reactions to PR006. However, Uspenskaya told Alzforum that this immunosuppressive cocktail appeared to have little effect in staving off CSF pleocytosis, and given the thrombosis risk, it will not be used in future trials. These serious adverse events did not correlate with increases in plasma or CSF progranulin.

One person died during the trial, from respiratory acidosis considered unrelated to treatment.

All told, the authors believe PR006 was reasonably well-tolerated and safe enough to warrant larger trials. Lilly is enrolling participants in a “bridging cohort” to evaluate a version of PR006 that is replicated in insect cells and can be scaled up for use in Phase 3 studies. Based on findings from both cohorts, an appropriate dose will be selected for larger trials, Uspenskaya said.

“One of the reasons we think this is a great approach is because it’s one and done,” Uspenskaya said. Still, she acknowledged the possibility that with time, the virus could be cleared from the brain. To check for this, progranulin levels will be tracked for five years in all participants in ongoing and future trials, she said. A second dose could be given, but then, the scientists may need to contend with immunity mounted against the vector, a general problem with AAV approaches.

Questions remain about biodistribution of CNS-injected drugs such as PR006. Sevigny and colleagues did not report the extent of the AAV-derived progranulin distribution in the brains of the macaques, though human progranulin RNA was found in three regions of the brain that were tested. Uspenskaya said that the drug reached the critical cortical regions of the brain most in need of progranulin replenishment in FTD. She noted that because the PR006-derived progranulin is secreted by transduced cells, not every cell needs to make it.

Meanwhile, a similar approach is in the works. Passage Bio’s PBFT02, an AAV1-PGRN construct injected into the cisterna magna, is being evaluated in a Phase 1/2 trial in people with FTD-GRN.—Jessica Shugart

News Citations

1. Pumping Up Progranulin: Scientists Show New Efforts to Get It Done 7 Apr 2023

Therapeutics Citations

1. PBFT02

Paper Citations

1. Logan T, Simon MJ, Rana A, Cherf GM, Srivastava A, Davis SS, Yoon Low RL, Chiu CL, Fang M, Huang F, Bhalla A, Llapashtica C, Prorok R, Pizzo ME, Calvert ME, Sun EW, Hsiao-Nakamoto J, Rajendra Y, Lexa KW, Srivastava DB, van Lengerich B, Wang J, Robles-Colmenares Y, Kim DJ, Duque J, Lenser M, Earr TK, Nguyen H, Chau R, Tsogtbaatar B, Ravi R, Skuja LL, Solanoy H, Rosen HJ, Boeve BF, Boxer AL, Heuer HW, Dennis MS, Kariolis MS, Monroe KM, Przybyla L, Sanchez PE, Meisner R, Diaz D, Henne KR, Watts RJ, Henry AG, Gunasekaran K, Astarita G, Suh JH, Lewcock JW, DeVos SL, Di Paolo G.
   Rescue of a lysosomal storage disorder caused by Grn loss of function with a brain penetrant progranulin biologic.
   Cell. 2024 Mar 14;187(6):1565-1566.
   PubMed.
2. Hordeaux J, Buza EL, Dyer C, Goode T, Mitchell TW, Richman L, Denton N, Hinderer C, Katz N, Schmid R, Miller R, Choudhury GR, Horiuchi M, Nambiar K, Yan H, Li M, Wilson JM.
   Adeno-Associated Virus-Induced Dorsal Root Ganglion Pathology.
   Hum Gene Ther. 2020 Aug;31(15-16):808-818. Epub 2020 Jul 31
   PubMed.

Papers

1. Amado DA, Rieders JM, Diatta F, Hernandez-Con P, Singer A, Mak JT, Zhang J, Lancaster E, Davidson BL, Chen-Plotkin AS.
   AAV-Mediated Progranulin Delivery to a Mouse Model of Progranulin Deficiency Causes T Cell-Mediated Toxicity.
   Mol Ther. 2019 Feb 6;27(2):465-478. Epub 2018 Nov 17
   PubMed.
2. Hinderer C, Miller R, Dyer C, Johansson J, Bell P, Buza E, Wilson JM.
   Adeno-associated virus serotype 1-based gene therapy for FTD caused by GRN mutations.
   Ann Clin Transl Neurol. 2020 Oct;7(10):1843-1853. Epub 2020 Sep 16
   PubMed.