Progranulin, a raw form of granulin (GRN), landed center stage of neurodegeneration research four years ago (see ARF related news story), when scientists discovered that loss-of-function mutations in the GRN gene cause a specific form of dementia—frontotemporal lobar degeneration with ubiquitin inclusions (FTLD-U). Fascination with the protein grew when studies later identified GRN missense mutations as potential risk factors for other neurodegenerative diseases, including Alzheimer’s (see GRN on AlzGene), Parkinson’s, and amyotrophic lateral sclerosis. But despite intense interest in progranulin, researchers have yet to discover what the protein does exactly. A major boost came last week with publications from two independent research groups who used completely different approaches to arrive at the same conclusion—that progranulin binds to the cell surface protein sortilin. “How this relates to disease is as yet unclear,” said Stephen Strittmatter, Yale University School of Medicine, New Haven, Connecticut, and senior author on one of the papers. “But it gives us something to study, as opposed to progranulin alone. Now we have a model we can get a handle on,” he told ARF.

Strittmatter’s group screened for progranulin binding partners using a brain tissue cDNA library. From among 200,000 clones they only found one, expressing sortilin, that bound with high affinity. Their work appeared in the November 18 Neuron. Rosa Rademakers at the Mayo Clinic College of Medicine, Jacksonville, Florida, led the second study. Rademakers and colleagues used a genomewide search to find genetic variants that could explain variability in plasma progranulin levels in healthy individuals. Their study, to be published in the December 10 American Journal of Human Genetics (currently available online), turned up two single-nucleotide polymorphisms (SNPs). The SNP with the most significant association (rs646776) lies close to the sortilin gene.

Rademakers and others previously found that plasma progranulin levels are lower in individuals with GRN loss-of-function mutations compared to healthy controls (see Finch et al., 2009 and ARF related news story). “But then we realized that there were huge, two- to threefold variations even in healthy individuals,” said Rademakers. To see if genetics could explain at least some of that variation, the researchers capitalized on the late-onset AD genomewide association study (GWAS) run by Steven Younkin at Mayo (see Carrasquillo et al., 2009). Joint first authors Minerva Carrasquillo, Alexandra Nicholson, and NiCole Finch and colleagues analyzed GWAS genotypes of 533 healthy Caucasian individuals and looked for a link between genetic variations and plasma progranulin. Two SNPs, both on chromosome 1p13.3, had positive associations. People carrying the minor allele (cytosine) of the most significant SNP, rs646776, had approximately 16 percent lower plasma progranulin than people with the major allele. The researchers replicated the finding using a different cohort of 508 healthy white controls. They also found that the rs646776 minor allele associated with lower plasma progranulin in a cohort of FTLD patients without GRN mutations. Furthermore, the presence of the minor allele seems to lower further the already low plasma progranulin in patients with GRN loss-of-function mutations. The findings point to a robust link between the chromosome 1 SNP and circulating progranulin.

How does the SNP exert its influence? Researchers recently showed that it elevates expression of three genes in the liver. They are CELSR2, a G-type receptor; PSRC1, a proline/serine rich protein; and SORT1, the gene for sortilin (see Kathiresan et al., 2008; Musunuru et al., 2010). Since the last is expressed on neurons, Rademakers wondered if it might be a key link between progranulin and neurodegeneration. She tested this by overexpressing or knocking down sortilin in HeLa cell cultures. In the former, there was a dearth of progranulin in the conditioned medium after five days compared to normal HeLa cells, while in the latter, progranulin levels in the medium rose almost twofold. Those findings suggest that sortilin expression is linked to trafficking or uptake of progranulin in cells.

Strittmatter’s Neuron paper supports this idea. Joint first authors Fenghua Hu, Thihan Padukkavidana, and colleagues used an alkaline phosphatase (AP)-tagged variant to detect cell surface receptors for progranulin. Finding that the AP-GRN bound to neurons, they then expressed the brain cDNA library in COS cells, which do not bind progranulin. From that screen, they pulled out sortilin and went on to show that it binds to progranulin in whole brain immunoprecipitates, and that binding of progranulin to neurons is almost nil if sortilin is knocked out. “That the binding is not completely abolished suggests that there could be some other progranulin binding partners in neurons,” said Strittmatter. What those other partners might be is unclear. Hu and colleagues tested other sortilin family members, including SorLA (see ARF related news story) and SorCS1 (see ARF related news story), which have both been implicated in amyloid precursor protein sorting and AD risk, but found that none bound to progranulin.

The consequences of the sortilin-progranulin interaction are also unclear. “Sortilin can be a signaling molecule, but we have not shown that with respect to progranulin,” said Strittmatter. When in cahoots with p75, sortilin mediates induction of apoptosis, or programmed cell death, driven by immature forms of neurotrophic factors, including proBDNF and proNGF (for a review see Al-Shawi et al., 2007). But Strittmatter noted that they did not detect such an outcome in their cell assay. Instead, sortilin may be acting as an endocytosis receptor. Hu and colleagues found that sortilin-positive COS cells rapidly take up added progranulin and deliver it to lysosomes. “This could be related to autophagy,” suggested Strittmatter, because the progranulin seems to stay associated with the lysosomal compartment, though, he added, that progranulin could be involved in signaling as well. “We really don’t know, yet,” said Strittmatter. Sam Gandy, Mount Sinai Medical School, New York City, weighed in on the complexities involved in an e-mail. “I see the sortilins as gateways to an enormously complex method of regulation,” he wrote (see full comment below). Gandy recently reported that the sortilin SorCS1 reduces amyloid-β production from APP (see ARF related news story).

“Whatever progranulin is doing, it needs to be linked somehow to accumulating TDP-43 and ubiquitin,” said Strittmatter. Inclusions that contain these proteins are hallmarks of FTLD-U. “Though we have provided a step, we have not linked progranulin to TDP-43 mishandling,” he said. His work does show 2.5-fold increased progranulin levels in the brains of sortilin-negative mice compared to wild-type. Since some familial FTLD-U is due to progranulin insufficiency, absence of sortilin could normalize progranulin levels, suggest the authors. “Another therapeutic implication is that PGRN-mediated signaling through sortilin could play a role in FTLD-U; if this is the case, then restoring that signaling could be a tractable therapeutic strategy,” suggest Jada Lewis and Todd Golde, University of Florida, Gainesville, in a Neuron Preview.

For their part, Rademakers and colleagues are still looking for other genetic variants that might give some clue as to what goes on in the FTLD brain. Though the SNP they identified does not affect sortilin expression in the brain, there could be others that do. She also plans to look for genetic variants that affect brain or cerebrospinal fluid levels of progranulin.—Tom Fagan


  1. I see the sortilins as gateways to an enormously complex method of regulation. Various members of the sortilin family can interact with either proneurotrophins or neurotrophin receptors and determine life-versus-death pathways. SorCS1/SorL1/APP complexes are in this mix, as are sortilin/progranulin interactions defined by Strittmatter. It is likely that multiple other family members can substitute for, or add onto, these complexes. To make matters even more complex, all the sortilins are γ-secretase substrates and undergo ectodomain shedding (usually by α-secretase). That may well convert them into diffusible ligands that activate yet undiscovered sortilin ectodomain receptors.

    To add one more layer of complexity, SorL1/SorLA/LR11 is a member of both the sortilin family and the LDLR/LRP family, so that regulation by lipoproteins might also pertain.

    It is somewhat surprising that members of the sortilin family are linked to both AD and FTD, and sortilin itself is a major risk for atherosclerosis (see Musunuru et al., 2010). One the one hand, this smacks of "too much good news"; on the other hand, the genetic linkages are solid, and the challenge now is to define the molecular neurobiology of sortilins and work out how they fit into the major human diseases of AD, FTD, diabetes, and atherosclerosis.


    . From noncoding variant to phenotype via SORT1 at the 1p13 cholesterol locus. Nature. 2010 Aug 5;466(7307):714-9. PubMed.

    View all comments by Sam Gandy

Make a Comment

To make a comment you must login or register.


News Citations

  1. Birds of a Feather…Mutations in Tau Gene Neighbor Progranulin Cause FTD
  2. Meet Progranulin, The Biomarker—A Simpler Story?
  3. Sorting Out SorLA—What Role in APP Processing, AD?
  4. APP Sorting Protein May Link Alzheimer’s and Diabetes

Paper Citations

  1. . Plasma progranulin levels predict progranulin mutation status in frontotemporal dementia patients and asymptomatic family members. Brain. 2009 Mar;132(Pt 3):583-91. PubMed.
  2. . Genetic variation in PCDH11X is associated with susceptibility to late-onset Alzheimer's disease. Nat Genet. 2009 Feb;41(2):192-8. PubMed.
  3. . Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet. 2008 Feb;40(2):189-97. PubMed.
  4. . From noncoding variant to phenotype via SORT1 at the 1p13 cholesterol locus. Nature. 2010 Aug 5;466(7307):714-9. PubMed.
  5. . ProNGF, sortilin, and age-related neurodegeneration. Ann N Y Acad Sci. 2007 Nov;1119:208-15. PubMed.

External Citations

  1. GRN on AlzGene

Further Reading

Primary Papers

  1. . Genome-wide screen identifies rs646776 near sortilin as a regulator of progranulin levels in human plasma. Am J Hum Genet. 2010 Dec 10;87(6):890-7. PubMed.
  2. . Sortilin-mediated endocytosis determines levels of the frontotemporal dementia protein, progranulin. Neuron. 2010 Nov 18;68(4):654-67. PubMed.
  3. . Sorting out frontotemporal dementia?. Neuron. 2010 Nov 18;68(4):601-3. PubMed.