Do Alzheimer’s risk genes act as lone wolves, or do they run as a pack? A paper in the 28 September Nature Neuroscience suggests the latter. Researchers led by Philip De Jager and Elizabeth Bradshaw at Brigham and Women’s Hospital, Boston, examined the expression of six AD-associated immune genes in human monocytes taken from healthy people. They found three different linkages between particular risk variants. Most notably, a risk allele of the CD33 receptor pumped up levels of wild-type TREM2 on the cell surface. Meanwhile, in two papers in the September 15 Journal of Biological Chemistry, independent research groups reported that TREM2 binds ApoE, establishing a link between the strongest genetic risk factors for late-onset AD. Though it is still unclear how these different TREM2 interactions might affect pathology, researchers noted that the data will provide new hypotheses to test.

Commenters commended the approach taken by De Jager and Bradshaw. “I appreciated this unbiased technique for linking AD-associated alleles with particular endophenotypes,” said Guojun Bu at the Mayo Clinic in Jacksonville, Florida. He led one of the two studies reported in the JBC. Marco Colonna at Washington University School of Medicine in St. Louis said, “This is the first data showing that some genetic polymorphisms [from AD GWAS] may be related.” At the same time, researchers noted that the findings contradict some previous studies suggesting that TREM2 expression benefits the brain. More work will be needed to pin down exactly how this immune receptor contributes to AD, they said.

Genetic studies have now turned up numerous immune genes that associate with Alzheimer’s. Previously, De Jager and colleagues found that several of these variants affected global gene-expression profiles in isolated human monocytes (see May 2014 news). His group is one of four funded through the National Institute of Health’s Accelerating Medicines Partnership to explore the molecular networks that lead to AD (see Feb 2014 newsFeb 2015 conference news). 

The link between genetic variants and coordinated gene expression in monocytes hinted that some of these genes might work together in common pathways. To investigate this, first author Gail Chan analyzed monocytes isolated from the blood of 115 healthy young people participating in the Phenogenetic Project at Brigham and Women’s Hospital, as well as 61 older, cognitively healthy adults enrolled in the Harvard Aging Brain Study. The Phenogenetic Project collects DNA and blood from healthy young and middle-aged volunteers to try to correlate gene expression with function. The authors measured protein levels in the monocytes, instead of mRNA as they had previously. They correlated the presence of 26 AD-related single-nucleotide polymorphisms (SNPs) in multiple genes with levels of six proteins: TREM1, TREM2, TREML2, CD33, TYROBP, and PTK2B. All these genes are expressed in monocytes and implicated in AD. The authors turned up three associations, which they validated in an independent cohort of 50 Phenogenetic Project participants.

In one of these, a risk variant near the NME8 gene associated with higher levels of PTK2B protein. High PTK2B expression previously had been linked to AD. Second, a TREM1 risk allele associated with lower expression also correlated with higher TREM2. Intriguingly, this relationship showed up only in monocytes from the young cohort, suggesting the interaction changes with age. Third, the authors found the link between CD33 and TREM2, in which a single SNP in the former boosted expression of both proteins at the cell surface. The SNP did not affect TREM2 mRNA levels. The authors estimated that the CD33 SNP accounted for 14 percent of the variability in the amount of TREM2. When they suppressed CD33 signaling in monocytes with an antibody, they saw a drop in TREM2 on the cell surface, further supporting the relationship.

The data puzzled several commenters. Colonna noted that CD33 activates protein tyrosine phosphatases, which inhibit cell activity, whereas TREM2 switches on protein tyrosine kinases. “They elicit opposite signaling,” Colonna said. Steven Estus at the University of Kentucky, Lexington, agreed. “There is no obvious reason why the two receptors should be regulated in parallel,” he told Alzforum. One possibility is that it represents a cellular homeostasis mechanism, where the cell turns down both receptors in order to maintain a balance between the two types of signaling, Estus speculated. It is also unclear how the CD33 receptor might bump up TREM2 levels. CD33 has been reported to regulate other receptors by binding them via their sialic acid residues, Estus noted (see Ishida et al., 2014).

Commentators were also struck by the association of high TREM2 with an AD risk factor, which hints that the TREM2 receptor exacerbates pathology. To investigate this, the authors analyzed TREM2 expression in prefrontal cortex samples taken postmortem in the Religious Orders Study and the Memory and Aging Project, both conducted at the Rush University Medical Center, Chicago. They found that elevated TREM2 correlated with greater amyloid load and higher odds of having AD. The findings fit with a recent study by Bruce Lamb and Richard Ransohoff, then at the Lerner Research Institute and Lou Ruvo Center for Brain Health, Cleveland Clinic, Ohio, and Gary Landreth at Case Western Reserve University, Cleveland, who found less inflammation and pathology in an AD mouse model after knocking out TREM2 (see Dec 2014 conference newsJay et al., 2015). 

However, others have reported that TREM2 stimulates phagocytosis and correlates with plaque clearance from the brain (see Feb 2015 news). In addition, missense mutations in TREM2 that associate with AD and frontotemporal dementia prevent the receptor from reaching the cell surface, again suggesting that TREM2 function is beneficial (see Jul 2014 webinar). These rare mutations were not present in the healthy cohort De Jager and Bradshaw analyzed.

Researchers agreed that several differences between TREM2 studies could account for the contradictory findings with regard to AD pathology. For example, the studies used different mouse models of amyloidosis, different TREM2 knockouts, and different TREM2 antibodies. Most existing TREM2 antibodies have been poorly validated, Bu noted. In addition, De Jager and Bradshaw examined mostly monocytes, whereas many previous studies describing TREM2’s benefits have focused on microglia. In future work, the authors will compare brain and monocyte expression data in a cohort of 300 people.

“These studies tell us we should not jump to conclusions about TREM2 too quickly,” Bu told Alzforum. Final answers on how TREM2 affects pathology may have to wait for animal models of the R47H mutation, which roughly triples the risk of AD, he suggested. Such models are being generated by several groups, including his.

In the meantime, another clue to TREM2’s function comes from the recent papers from Bu’s lab and researchers led by Charles Bailey at the Scripps Research Institute in Jupiter, Florida. Each group reported that microglial TREM2 binds ApoE with high affinity in vitro, and that the R47H allele weakens this binding. Microglia and macrophages produce a lot of ApoE, and production shoots up after peripheral nerve injury, Bu noted. Because ApoE binds lipids, the lipoprotein may help phagocytes that express TREM2 clean up cellular debris present during degeneration, Bu suggested.

In line with this, first author Yuka Atagi in Bu’s group found that cultured microglia that bound ApoE through TREM2 chewed up nearby dying cells more efficiently. Interestingly, the ApoE4 risk allele did not affect binding to TREM2, but Bu noted that overall expression of the ApoE4 allele is lower than ApoE2/3, and that might impair phagocytosis. The findings dovetail with other work suggesting that TREM2 plays a key role in general brain cleanup, rather than amyloid clearance specifically (see Jun 2014 news; Apr 2015 conference news).—Madolyn Bowman Rogers


  1. The study by Chan et al. follows nicely from previous work of the same group focused on the dissection of the molecular mechanisms underlying the genetic loci associated with AD risk.

    These studies are essential because the majority of the genetic variants associated with AD risk in genome-wide association studies are located in intergenic or intronic regions and it has not been straightforward to understand how these loci are functionally related to the disease. One assumption was that many of these variants would be associated with differences in gene expression. Failure to validate this premise for many of the loci indicated the need for studies such as this one, which performs protein quantitative trait analysis in a subset of cells expected to be important, or to mimic cells with known involvement in the pathological process of the disease. 

    This evaluation of the relation between known AD risk loci showed that the NME8 risk allele influences PTK2B, the CD33 risk allele influences TREM2, and the TREM1 risk allele is associated with a decreased TREM1/TREM2 ratio. The role of TREM2 in AD is still unclear and results from studies like this one are essential to better understand how genetic variability at this locus contributes to disease. In this case the results obtained favor a pathogenic role of increased TREM2 expression in myeloid cells from the periphery. 


  2. This article by Chan and colleagues offers new insights into the genetics and pathogenic effects of CD33 and TREM2 in Alzheimer’s disease (AD). CD33 was first implicated in AD in 2008, when we reported that the minor allele (G) of the rs3826656 single nucleotide polymorphism (SNP) in the CD33 gene confers risk for late-onset AD (Bertram et al., 2008). CD33 is a sialic acid-binding protein that we recently found to be expressed by microglial cells in the human aging brain (Griciuc et al., 2013). We had previously reported in vivo and in vitro evidence that CD33 inhibits microglia-mediated clearance of Aβ and promotes the formation of amyloid plaques in the brain (Griciuc et al., 2013). In the same study, we showed that the minor allele (A) of another CD33 SNP rs3865444, which confers protection against AD, was associated with reductions in both CD33 expression levels in microglial cells and insoluble Aβ42 levels in AD brain (Griciuc et al., 2013). In another study, the major allele (C) of rs3865444 was associated with increased CD33 expression levels in peripheral monocytes (Bradshaw et al., 2013). The molecular mechanisms underlying the effects of the rs3865444 and rs3826656 SNPs on AD risk have remained unclear.

    In their article, Chan et al. measured protein levels on the surface of peripheral monocytes from blood samples collected from 115 young subjects and 61 cognitively-intact aged individuals. The authors found that rs3865444C and rs3826656A alleles were associated with increased CD33 surface expression on peripheral monocytes.  However, rs3865444C and rs3826656A result in contrasting associations with susceptibility to late-onset AD. Therefore, it will be important to investigate whether the rs3826656A allele is also associated with increased CD33 levels in microglial cells in cognitively intact and AD brains, and how it impacts CD33 isoforms.

    Interestingly, Chan and colleagues found that the presence of the rs3865444C in CD33 was also associated with increased expression levels of TREM2 on the surface of peripheral monocytes. Further, they showed that suppressing CD33 signaling with a CD33-specific antibody resulted in decreased TREM2 surface expression levels in monocytes. It therefore will be important in future studies to dissect the molecular mechanisms underlying the cross-regulation between CD33 and TREM2 in peripheral monocytes and microglial cells, as well as in healthy aged versus AD brains. 

    Rare mutations in the TREM2 gene result in susceptibility to late-onset AD (Lill et al., 2015). Colonna and colleagues showed that TREM2 plays a beneficial role in AD by sustaining the microglial response to amyloid beta accumulation (Wang et al., 2015; Tanzi, 2015). In contrast, Jay et al. found that TREM2-positive brain macrophages play a detrimental role in mouse models of AD (Tanzi, 2015; Jay et al., 2015). In the new study, Chan et al. found that increased TREM2 mRNA levels were associated with greater amyloid load in the human prefrontal cortex, thus suggesting that increased TREM2 levels play a pathogenic role and lead to AD susceptibility. Therefore, it remains unclear whether increased or decreased TREM2 expression levels increase risk for late-onset AD (Tanzi, 2015; Jay et al., 2015). Future studies will be needed to resolve these discrepant results regarding the role of TREM2 in AD pathology. Meanwhile, the majority of the available data suggest that inhibiting CD33 remains a potentially useful strategy for ameliorating AD pathology.


    . Genome-wide association analysis reveals putative Alzheimer's disease susceptibility loci in addition to APOE. Am J Hum Genet. 2008 Nov;83(5):623-32. PubMed.

    . Alzheimer's disease risk gene CD33 inhibits microglial uptake of amyloid beta. Neuron. 2013 May 22;78(4):631-43. PubMed.

    . CD33 Alzheimer's disease locus: altered monocyte function and amyloid biology. Nat Neurosci. 2013 Jul;16(7):848-50. PubMed.

    . The role of TREM2 R47H as a risk factor for Alzheimer's disease, frontotemporal lobar degeneration, amyotrophic lateral sclerosis, and Parkinson's disease. Alzheimers Dement. 2015 Dec;11(12):1407-1416. Epub 2015 Apr 30 PubMed.

    . TREM2 lipid sensing sustains the microglial response in an Alzheimer's disease model. Cell. 2015 Mar 12;160(6):1061-71. Epub 2015 Feb 26 PubMed.

    . TREM2 and Risk of Alzheimer's Disease--Friend or Foe?. N Engl J Med. 2015 Jun 25;372(26):2564-5. PubMed.

    . TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer's disease mouse models. J Exp Med. 2015 Mar 9;212(3):287-95. Epub 2015 Mar 2 PubMed.

Make a Comment

To make a comment you must login or register.


Alzpedia Citations

  1. CD33
  2. TREM2

News Citations

  1. Alzheimer's GWAS Hits Reflected in Monocyte Gene Expression
  2. New Initiative AMPs Up Alzheimer’s Research
  3. Alzheimer's Disease Research Summit 2015: Expanding the Horizon
  4. TREM2 Data Surprise at SfN Annual Meeting
  5. TREM2 Buoys Microglial Disaster Relief Efforts in AD and Stroke
  6. TREM2 Mystery: Altered Microglia, No Effect on Plaques
  7. Microglia—Who Are You Really? New Clues Emerge

Webinar Citations

  1. Mutations Impair TREM2 Maturation, Processing, and Microglial Phagocytosis

Paper Citations

  1. . Negative regulation of Toll-like receptor-4 signaling through the binding of glycosylphosphatidylinositol-anchored glycoprotein, CD14, with the sialic acid-binding lectin, CD33. J Biol Chem. 2014 Sep 5;289(36):25341-50. Epub 2014 Jul 24 PubMed.
  2. . TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer's disease mouse models. J Exp Med. 2015 Mar 9;212(3):287-95. Epub 2015 Mar 2 PubMed.

External Citations

  1. Phenogenetic Project

Further Reading

Primary Papers

  1. . CD33 modulates TREM2: convergence of Alzheimer loci. Nat Neurosci. 2015 Nov;18(11):1556-8. Epub 2015 Sep 28 PubMed.
  2. . Apolipoprotein E Is a Ligand for Triggering Receptor Expressed on Myeloid Cells 2 (TREM2). J Biol Chem. 2015 Oct 23;290(43):26043-50. Epub 2015 Sep 15 PubMed.
  3. . The Triggering Receptor Expressed on Myeloid Cells 2 Binds Apolipoprotein E. J Biol Chem. 2015 Oct 23;290(43):26033-42. Epub 2015 Sep 15 PubMed.