In the blood, lipoproteins ferry cholesterol and other lipids to cells throughout the body. In the cerebrospinal fluid, they may do the same, but because they are 200-fold more dilute there, scientists know little about them. Now, John Melchior at the Pacific Northwest National Laboratory in Richland, Washington, and colleagues have overcome this hurdle. They discovered a wide array of lipoproteins in the CSF, which even surpasses that of blood. In the August 31 Science Translational Medicine, the scientists reported 10 differently sized classes of CSF lipoproteins for a total of 303. These lipoproteins are involved in processes germane to neurodegeneration, including inflammation, immune responses, wound healing, and synapse function.

  • Ten size-groups of lipoproteins cruise healthy cerebrospinal fluid.
  • They comprise 303 proteins, of which half are not in plasma.
  • Their pathways? Inflammation, immune response, and more.

Other scientists praised the work. “[It] represents a significant advancement in our understanding of which proteins are present in human CSF lipoprotein particles,” wrote Tobias Hartmann, Saarland University in Homburg, Germany. René Frank, University of Leeds, U.K., noted that some of the proteins in these particles have been associated with Alzheimer’s disease. “[This study is] a treasure trove of biochemical information on the protein composition of cognitively normal human donor CSF and the size profile of [its] lipoprotein particles and extracellular vesicles,” he wrote. “... It could be exciting to extend this elegant study of human CSF to that from living Alzheimer’s disease patient donors, comparing with CSF from healthy, and cognitively normal, aged-matched control donors.” (Comments below.)

Plasma lipoproteins come in four flavors: ultra-low-density, very-low-density (VLDL), low-density (LDL), and high-density (HDL). Lipoproteins include, you guessed it, fats—phospholipids surrounding a core of cholesterol and triglycerides—and proteins, commonly apolipoproteins. Their makeup varies widely, with HDLs housing up to 285 types of protein (reviewed by Davidson et al., 2022). Until now, the CSF lipoprotein list had a measly six, all apolipoproteins.

To plumb the CSF lipoproteome, first author Nathaniel Merrill analyzed commercially available pooled human CSF samples. He added fluorescent lipids to label the lipoproteins, which made them more detectable. Then, he used high-resolution size-exclusion chromatography (SEC) to separate them through three tandem columns, followed by mass spectrometry to identify the proteins (see image below).

Voila, Lipoproteins. After labeling them with a fluorescent lipid, lipoproteins were separated by size exclusion chromatography. A lipid removal agent (LRA) left just the proteins, which were then digested with trypsin and analyzed via mass spectrometry. [Courtesy of Merrill et al., Science Advances, 2023.]

Merrill detected 10 different-sized clusters of lipoproteins: one large VLDL and one large LDL, seven small-to-medium HDLs, and one very small HDL. Of the 303 proteins, half had not been detected in blood lipoprotein particles. The more abundant of these included the serine protease kallikrein-6, neurotransmitter-related secretogranin-1, and neuropeptide precursor VGF. CSF-specific proteins accounted for just 15 percent of the total lipoprotein content. The most abundant lipoproteins in CSF were also found in blood, e.g. APOA1, APOA2, APOE, and APOJ, aka clusterin.

To the authors, this means the CSF-specific proteins might dock onto lipoprotein particles to give them specific functions. These proteins eluted from the SEC column in single peaks, indicating they hitched a ride with only one class of lipoprotein. At the same time, apolipoproteins acted as the primary scaffolds for forming the particles, eluting with most lipoprotein species in the samples.

Lipoprotein Configuration. Different-sized lipoproteins contain unique combinations of proteins, including ones specific to CSF (red) or also found in blood (gray). This diagram shows a sample of 157 such protein groups identified. [Courtesy of Merrill et al., Science Advances, 2023.]

What might these proteins be doing in the brain? Gene ontology placed them into 15 functional clusters, including immune responses, inflammation, wound healing, proteolysis, and neurodevelopment. Many CSF-specific proteins were in the latter group. Machine-learning algorithms pieced together protein-protein relationships, finding about 150 groups of two to four that likely hitch rides on the same-sized lipoprotein (see image below).

One group of proteins caught the researchers’ attention. It involved neurexin 3 (NRXN3), amyloid precursor-like-protein 1 (APLP1), and plexin domain containing 2 (PLXDC2). All are substrates of β-secretase, one of two enzymes needed to cut Aβ from APP. NRXN3 and APLP1 also regulate synapse formation and function (Aoto et al., 2015; Schilling et al., 2017).

Commentators wondered how CSF lipoprotein composition changes with aging and during neurodegenerative diseases, such as AD (comments below). “It will be critical to profile changes of the CSF lipoproteome during the course of Alzheimer’s disease progression, particularly in individuals at high risk for AD, such as carriers of autosomal dominant AD mutations or of APOE4/4, and those carrying loss-of-function mutations in AD risk genes implicated in lipoprotein metabolism,” wrote Edoardo Marcora, Icahn School of Medicine, Mount Sinai, New York.—Chelsea Weidman Burke


  1. This work represents a significant advancement in our understanding of which proteins are present in human CSF lipoprotein particles, the brain's counterpart to LDL and HDL. Yesterday, there were a mere dozen or so proteins identified; today, that number exceeds 300. Although the authors employ a relatively broad definition of which proteins qualify, and all the newly identified proteins are categorized as less abundant, this research sheds profound light on the brain’s lipid-protein interactions and their functions. The precise implications of this discovery will undoubtedly necessitate years of intricate research. I’m especially intrigued by the preliminary finding that there might be notable and comprehensive variations from one patient to another. Another compelling facet of this study is its suggestion that CSF lipoprotein particles can be classified into 10 or more distinct-size clusters, which implies the existence of functional clusters—which, interestingly, may also exhibit inter-individual variations.

  2. Merrill et al. report a treasure drove of biochemical information on the protein composition of cognitively normal human donor CSF and the size profile of lipid-containing constituents (lipoprotein particles and extracellular vesicles) of CSF. Of particular note, lipid-containing CSF constituents were found to be separable into 10 distinct types based on size, as per retention volume in biochemical SEC experiments.

    Distinct types were enrichment with particular proteins, including some that are associated with Alzheimer’s disease. I wonder to what extent these CSF profiles vary, if at all, with age, genetic and environmental factors, e.g., diet/cardiovascular health?

    In the future, it could be exciting to extend this elegant study of human CSF to that from living Alzheimer’s disease patient donors compared to healthy and cognitively normal aged-matched control donors. 

  3. Both old and new human genetic findings strongly implicate lipoprotein metabolism in the etiology of AD. For example, apolipoprotein genes APOE and CLU/APOJ, and ABC transporter genes ABCA1 and ABCA7, which load lipids onto apolipoproteins to produce lipoproteins, all have been associated to AD by GWAS (or even before GWAS in the case of APOE). In addition, given that the brain is the most lipid-rich organ in the human body, lipoprotein metabolism is likely to be important for its normal development and functioning.

    However, the few humans identified with loss-of-function variants in APOE have grossly normal brains and may even be resistant to AD pathology (Chemparathy et al., 2023). These seemingly paradoxical findings highlight our ignorance of lipoprotein metabolism in the brain and of its relationship to brain development, homeostasis, aging, and disease. Therefore, the novel methods and findings from studies such as this one are necessary to improve our understanding of such a fundamental biological system and its role in AD pathogenesis.

    For instance, the authors show that the compositional and functional complexity of CSF lipoproteins is likely to rival that of plasma high-density lipoproteins, which have been implicated in several functions besides lipid transport, including the innate immune response. Importantly, lipid transport and the innate immune response may represent two sides of the same coin (Seong and Matzinger, 2004), and both pathways are strongly implicated in AD pathogenesis by GWAS.

    It will be critical to profile changes of the CSF lipoproteome during the course of AD progression, from early preclinical to late clinical stages. This is especially true for individuals at high risk (e.g., autosomal-dominant AD mutation or APOE4/4 carriers) and those carrying loss-of-function mutations in AD risk genes implicated in lipoprotein metabolism (e.g., APOE, CLU/APOJ, ABCA1, and ABCA7).

    Deeper investigations of the biological functions of lipoproteins in the CSF and in the interstitial fluid, beyond their role in amyloid deposition, e.g., in maintaining brain cholesterol homeostasis during aging and demyelination, or in the efferocytosis of lipid-rich cellular debris, such as apoptotic synapses by microglia (Romero-Molina et al, 2023), will likely be the key to a more complete understanding of AD pathobiology.


    . APOE loss-of-function variants: Compatible with longevity and associated with resistance to Alzheimer's Disease pathology. medRxiv. 2023 Jul 24; PubMed.

    . Hydrophobicity: an ancient damage-associated molecular pattern that initiates innate immune responses. Nat Rev Immunol. 2004 Jun;4(6):469-78. PubMed.

    . Microglial efferocytosis: Diving into the Alzheimer's disease gene pool. Neuron. 2022 Nov 2;110(21):3513-3533. PubMed.

  4. The study of plasma lipoproteins is a field unto itself, with decades of work profiling the variety, heterogeneity, protein/lipid content, and risk profile of these peripheral particles. When it comes to the brain, previous work by Weisgraber, Mahley, Yassine, and many other groups has shed some light on CSF lipoprotein classes, particularly in regard to one of the most abundant CSF apolipoproteins, ApoE, as well as the relationship between CSF lipoproteins and AD risk. However, the field has been limited by an issue of abundance, as CSF lipoproteins are generally found at levels two orders of magnitude lower than in plasma. For this reason, this paper is very exciting as it allows for a much deeper study of CSF lipoproteins. It opens up new opportunities to examine these particles across a number of neurodegenerative settings, including Alzheimer’s disease.

    We find the incorporation of a fluorescent lipid tracer to improve the sensitivity of size profiling by size-exclusion chromatography to be quite innovative. The deconvolution approach to identify the 10 sub-populations of CSF lipoproteins was also exciting. To our knowledge, this is by far the highest-resolution characterization of CSF lipoproteins published to date. Aside from the technical developments, one big takeaway was that about half of the proteins identified on the CSF lipoproteins (“HDL-like”) have not been reported to associate with plasma HDL, suggesting that these brain-derived particles have a rather unique protein composition.

    The implications of this on the field of lipoprotein function are very exciting and present a wealth of new opportunities for discovery of novel mechanisms by which protein-lipid interactions may impact physiology in health and disease.

    For those interested in digging into the full list of proteins identified, the authors provide a supplemental table, S1, to search for your favorite protein of interest. They also link to a resource that is likely of interest to others in the field, the Brain Lipoprotein Proteome Watch. Hosted by the lab of Sean Davidson and Amy Shah, this webpage contains a link to a downloadable, easily navigable, Excel file updated by Melchior. It consolidates and organizes the nine peer-reviewed proteomics studies that examined lipoproteins isolated from human CSF and made those proteomics data freely available to the research community. Browsing this table, it’s immediately clear how many more CSF lipoprotein-associated proteins were able to be detected using this new technique, as the vast majority were observed only by the current study (and not the eight prior).

    Several of the proteins identified on CSF lipoproteins in the current study may catch the eye of AD researchers—for example APOE, CLU, C3, or SPP1.

    Regarding APOE, the authors show that APOE co-elutes with the bulk of lipid. This, together with previous studies, suggests that APOE is the primary scaffolding protein of CSF lipoproteins—and we would have loved to see the authors try to develop an APOE protein network.

    Future studies that validate the proposed interaction models, as well as studies looking at how enrichment of these proteins on CSF particles (particularly those previously implicated in AD pathophysiology) changes during aging and disease, will be very interesting to follow.

  5. This work describing the diversity of human CSF lipoprotein species is timely and likely to have a strong impact on the field of Alzheimer’s disease.

    Previous elegant, yet perhaps underappreciated, studies have described differences in CSF particles in AD, including differences in size and lipid content. For example, Fonteh et al. reported altered fatty acid content in brain-derived nanoparticle fractions in both mild cognitive impairment and AD compared to healthy controls (Fonteh et al., 2014). The next year, Yang et al. reported differences in concentration of ApoAI-positive particles from CSF of those with MCI and AD. Interestingly, these particles also had differential annexin V positivity indicating altered phosphatidylserine content (Yang et al., 2015). Yang et al. also reported a difference in the number and size of apoE-positive particles from CSF of APOE e4/e4 homozygous study participants. The number and size of high-density lipoprotein (HDL) particles in blood of has been proposed to be predictive of cardiovascular disease (Kontush, 2015) supporting the promise of these endogenous nanoparticles (eNP) as potential biomarkers in AD. In fact, Martinez et al., and Hussein Yassine’s group at USC have proposed that small HDL particles, which they found were positively associated with cognitive function, may be protective due to capacity for lipid exchange (Martinez et al., 2022). 

    These eNP are also coming to the fore based on the diagnostic propensity of the NMR-based biomarker profile of the Nightingale Health NMR platform (Julkunen et al., 2023). It includes lipoprotein lipids identified by apolipoproteins, subclasses, size, and lipid content. However, the specific lipid and protein content of these eNP has yet to be determined. Further studies are critical to understand the specificity and function of the eNP, as well as their value as diagnostic or prognostic biomarkers.

    Merrill et al. use an innovative method for fluorescently labeling the eNP from CSF, CSF-lipoproteins (CSF-lps), allowing for separation by size exclusion chromatography. They elegantly validate the use of Liss Rhodamine phosphoethanolamine, which does not change the lipoprotein size or selectively label a subset of the CSF-lipoproteins (CSF-lps) but allows for the detection of the full cadre of apolipoproteins in CSF. This method facilitates the identification and speciation of CSF-lps using three size-exclusion columns in series with 350 μl CSF—a considerable improvement in the sensitivity for detection of 157 novel CSF-lps-associated proteins and an additional 146 which are common to HDL.

    Using the DAVID functional enrichment computational approach, Merrill et al. identified association of the proteome with several novel functions, including axon cell growth (neuronal cell adhesion molecule [NRCAM] and contactins 4 and 6) and substrates of β-secretase 1 (neurexin 3, amyloid precursor-like-protein1 and plexin domain containing 2), a protease relevant to AD cleavage of Aβ-peptide. They also identified serotransferrin and hemopexin, which have been shown in plasma to be associated with HDL and Aβ metabolism in AD (see the paper’s discussion and references).

    This paper strongly supports emerging evidence that CSF-derived lipoproteins are diverse, contribute to previously unrecognized basic biological functions, and are culpable in Alzheimer’s disease. Further study regarding function and specific dysregulation of different eNP species is required to fully understand the contribution of eNP dysregulation in disease states, which may lead to utility as biomarkers or potential therapeutic targets.


    . Human cerebrospinal fluid fatty acid levels differ between supernatant fluid and brain-derived nanoparticle fractions, and are altered in Alzheimer's disease. PLoS One. 2014;9(6):e100519. Epub 2014 Jun 23 PubMed.

    . Cerebrospinal Fluid Particles in Alzheimer Disease and Parkinson Disease. J Neuropathol Exp Neurol. 2015 Jul;74(7):672-87. PubMed.

    . The small HDL particle hypothesis of Alzheimer's disease. Alzheimers Dement. 2022 Apr 13; PubMed.

    . HDL particle number and size as predictors of cardiovascular disease. Front Pharmacol. 2015;6:218. Epub 2015 Oct 5 PubMed.

    . Atlas of plasma NMR biomarkers for health and disease in 118,461 individuals from the UK Biobank. Nat Commun. 2023 Feb 3;14(1):604. PubMed.

  6. This very interesting study focuses on human CSF functional lipoproteins by identifying novel proteins and networks specific to CSF that are involved in multiple CNS diseases. Especially novel is the experimental design of lipid size-based isolation, followed by proteomics to build lipoprotein interaction networks and identify proteins and regulating pathways unique to CSF compared to plasma.

    This study gives us insight into what proteins are involved in CSF lipoproteins and how these lipoproteins are associated with CNS development and regulations to maintain cellular homeostasis of both neuronal and glial health. We also learn about how these lipoproteins, potentially derived from different CNS cell types, can be involved in neurodegeneration and inflammation.

  7. Merrill et al. uncover the remarkable heterogeneity and brain-specific nature of lipoprotein particles (Lps) in the cerebrospinal fluid (CSF) of the human brain. To date, such CSF-Lps have remained understudied in comparison to the analogous lipoprotein particles found in blood plasma, which serve broadly similar roles in lipid transport and have previously been shown to exhibit a high degree of compositional and structural complexity.

    Prior in-depth analysis of CSF-Lps has been limited due to their low abundance (~0.5 percent the level of Lps in blood plasma) in the volume-limited clinical CSF. To overcome these technical challenges, the authors incorporate a fluorophore-tagged lipid to enable sensitive, low-volume fractionation of CSF-Lps via size-exclusion chromatography (SEC) coupled to liquid chromatography with tandem mass spectrometry (LC-MS/MS) for protein identification. Interestingly, 10 distinctly sized populations of CSF-Lps are detected, wherein most are larger in hydrodynamic size than plasma high-density lipoprotein (HDL; previously considered most similar to CSF-Lps based on size). Moreover, over half of the approximately 300 identified proteins are seemingly unique to CSF-Lps compared to plasma HDL, with proteins involved across various neurometabolic functions.

    Computational and bioinformatic analysis reveals that CSF-Lps contain proteins involved in immune response, inflammation, wound healing, and neuronal generation and development. The latter distinguishes these lipoprotein particles from those of plasma. Although the primary scaffolding proteins of CSF-Lps (including ApoE, ApoJ, and ApoA-I) are found to associate with CSF-Lps across a broad range of sizes, most other proteins demonstrate preference to reside on CSF-Lps of specific sizes. This speciation suggests that subpopulations of CSF-Lps may be defined by distinct protein fingerprints that impart discrete biological functions. They merit further investigation.

    More broadly, this lipoprotein phenotyping pipeline opens doors to exploring how CSF-Lps subpopulations of different protein compositional signatures may be implicated in maintaining CNS health and/or driving neurodegenerative disease states associated with aberrant lipid metabolism.

  8. Lipoproteins are circulating particles whose primary function is to carry hydrophobic lipids in aqueous environment. Recent evidence has unveiled the intricate nature of lipoproteins. In particular, several groups found that the high-density lipoproteins (HDL) within the plasma circulation are composed of over 200 lipid and 250 protein species.

    With the exception of specific apolipoproteins such as apoA-I that form the core of the lipoproteins, the proteins and lipids constituting HDL are distributed between different subpopulations. Given that these specific proteins and lipids determine a particle’s function and can be altered in diseases (Sacks et al., 2022), it is crucial to precisely and sensitively define its composition.

    While it was known that lipoproteins in cerebrospinal fluid (CSF) are of similar densities as plasma HDL, their composition remained enigmatic, partly due to the technical constraints arising from lower concentration of the lipoproteins in CSF compared to plasma.

    As reported in this elegant study, Merrill and colleagues developed a novel method to characterize CSF lipoproteins. Firstly, they labeled the lipids fluorescently and fractionated the lipoproteins using size-exclusion chromatography, coupled with an ultrasensitive fluorescent detector. Secondly, they analyzed the proteome composition of the lipid-containing particles using LC-MS/MS.

    Through size exclusion chromatography, they observed that CSF lipoproteins might form 10 distinct size populations, with most being similar in size or slightly larger in size than large plasma HDL. However, they also identified particles as large as LDL (peak 1 and 2) or much smaller than any plasma lipoproteins (peak 10). Using LC-MS/MS they identified more than 303 proteins constituting CSF lipoproteins, with only half of them present in plasma HDL.

    Computational analysis was used to identify 15 biological pathways associated with CSF lipoprotein proteome. Several, like inflammation and wound healing, are similar to plasma HDL, while others, like central nervous development or regulation, are specific to the brain. Finally, analyses of different fractions revealed unique proteome signatures for each lipoprotein subpopulation.

    Together, these results shed new light on the complexity of CSF lipoproteins and their potential roles in brain health. Particularly intriguing is the observation that a significant proportion of CSF lipoproteins resemble plasma HDL, implying transport across the CSF barrier. We previously reported that lipid-free apoA-I is transported through the choroid plexus endothelial cells (Stukas et al., 2014). However when HDL was injected into mice, it was predominantly found in the microvasculature (Sulliman et al., 2021). The mechanisms by which HDL enters the CSF remain to be investigated.

    Another proportion of the CSF lipoproteins differ from plasma HDL, suggesting production within the central nervous system. How these two pools of lipoproteins interact remains an open question.

    It is striking to observe inter-individual variability despite low inter-assay variation. Studies involving larger cohorts of individuals, both healthy and those with neurodegenerative diseases, may eventually contribute to biomarker identification.

    Given that ApoE is the primary genetic risk factor for Alzheimer’s disease, future investigation should explore how APOE genotype influences CSF lipoprotein composition. Previous work indicated that apoE4 lipoproteins are larger (Heinsinger et al., 2016), but their protein signatures await further investigations. Similarly, CSF levels of apoA-I, apoJ, or apoB are associated with Alzheimer’s disease (Picard et al., 2021). It remains to be determined whether these associations are solely due to altered levels or if they are related to distinct lipoprotein subpopulation signatures. 

    It is important to note that out of the 941 proteins identified by various research groups working on plasma HDL, “only” 250 were confirmed by at least three independent groups. These disparities may be attributed to different methods of isolation and measurement. The work of Merrill and colleagues, therefore, lays the foundation for future studies characterizing CSF lipoproteins. We should acknowledge the group initiative to compile existing and forthcoming data on the Brain Lipoprotein Proteome Watch.



    . Protein-based HDL subspecies: Rationale and association with cardiovascular disease, diabetes, stroke, and dementia. Biochim Biophys Acta Mol Cell Biol Lipids. 2022 Sep;1867(9):159182. Epub 2022 May 20 PubMed.

    . Intravenously injected human apolipoprotein A-I rapidly enters the central nervous system via the choroid plexus. J Am Heart Assoc. 2014 Nov 12;3(6):e001156. PubMed.

    . HDL biodistribution and brain receptors in zebrafish, using HDLs as vectors for targeting endothelial cells and neural progenitors. Sci Rep. 2021 Mar 19;11(1):6439. PubMed.

    . Apolipoprotein E Genotype Affects Size of ApoE Complexes in Cerebrospinal Fluid. J Neuropathol Exp Neurol. 2016 Oct;75(10):918-924. Epub 2016 Aug 11 PubMed.

    . Apolipoprotein B is a novel marker for early tau pathology in Alzheimer's disease. Alzheimers Dement. 2021 Sep 29; PubMed.

  9. Merrill et al. developed a method to detect and characterize cerebrospinal fluid (CSF) lipoproteins by doping in small amounts of fluorescent lipids for enhanced detection and using tandem size exclusion columns to separate a large range of particle sizes. The authors characterize the proteome of these different lipoprotein particles using mass spectrometry. This is an extremely powerful technique as it allows for the characterization of patient samples with medically tractable amounts of biological material.

    CSF lipoproteins adopted a different size distribution and distinct proteome from plasma lipoproteins. Although we’ve known that CSF lipoproteins and HDL lipoprotein abundance is qualitatively and quantitatively different (Pitas et al., 1987), this study quantifies the sizes of lipoprotein particles from human CSF samples. Multiple studies have demonstrated that different lipoprotein particles have distinct functional effects in models and in patients (Koch et al., 2020; Robert et al., 2020). The techniques in this study can be used to associate protein and lipid differences among lipoprotein particles with their functional roles in development and disease. These techniques could be extended to understanding dynamics of CSF lipoprotein particle size and composition in various developmental and disease contexts.

    Given that CSF lipoprotein particles are far less abundant than plasma lipoprotein particles, the approach described here enables sampling and characterization of previously intractable species. With this hurdle overcome, researchers can use this approach to associate various lipoprotein particles with risk, resilience, or staging of various CNS diseases. These sorts of studies will form the foundation for using CSF-lipoprotein particles as diagnostics or biomarkers for disease states akin to the use of plasma HDL and LDL in cardiovascular disease.

    The work from Merrill et al. has provided a valuable technique to the lipoprotein particle research community and opened the doors to a greater molecular and functional understanding of CSF lipoprotein particles in the context of health and disease.


    . Association of Apolipoprotein E in Lipoprotein Subspecies With Risk of Dementia. JAMA Netw Open. 2020 Jul 1;3(7):e209250. PubMed.

    . Lipoproteins and their receptors in the central nervous system. Characterization of the lipoproteins in cerebrospinal fluid and identification of apolipoprotein B,E(LDL) receptors in the brain. J Biol Chem. 1987 Oct 15;262(29):14352-60. PubMed.

    . Cerebrovascular amyloid Angiopathy in bioengineered vessels is reduced by high-density lipoprotein particles enriched in Apolipoprotein E. Mol Neurodegener. 2020 Mar 25;15(1):23. PubMed.

  10. This exciting piece of technical work makes use of state-of-the art analytical technologies including fluorescence-activated SEC and mass spectrometry-based proteomics. The recent developments within proteomics and masspec with respect to sensitivity and microscale preparation have transformed proteomics toward single-cell analysis (as in individual cells, not cell types).

    The downscaled and robust setup requires only relatively minute amounts of CSF, i.e. 350 microliters. This rather technical work paves the way for upscaled CSF-Lps profiling, which should now be validated in larger sample sets for both robust quantification and, ultimately, application to relevant clinical cohorts.

  11. The key contribution of this work is the description of the novel method combining fluorescent phospholipid labeling with tandem size-exclusion chromatography to investigate the composition of CSF lipoproteins at a high resolution. LC-MS/MS proteomic analysis of the isolated lipoparticles also enabled the identification of well-known CSF lipoprotein-associated proteins, such as ApoE, ApoAI, complement factors, and adhesion proteins. Importantly, this analysis revealed novel CSF lipoparticle proteins potentially involved in central nervous system regulation.

    It is important to note that the authors acknowledged a potential caveat in the form of lipid vesicle contamination lacking lipoproteins. Such binding to vesicles could be more general than specific to lipoproteins. Another small but significant advantage is the online program they designed to combine data on CSF lipoprotein proteins discovered by multiple research groups, which will facilitate cross-study comparisons.

    Critically, in this publication the authors restricted their analysis of CSF lipoproteins to only healthy subjects. In future studies, for example, given the significance of ApoE polymorphism in late-onset AD, PDD, LBD, and related dementias, this enhanced sensitivity and resolution tool could greatly contribute to our understanding of CSF lipoprotein differences among ApoE4 carriers, ApoE3 carriers, and ApoE2 carriers, where proper characterization is currently lacking due to limitations associated with the CSF matrix. Discovering variations in CSF lipoprotein composition in patients affected by the aforementioned diseases could yield potential new disease biomarkers.

    In the CSF lipoprotein proteome, the authors identified noteworthy new members are the APP and amyloid precursor-like proteins. Given the relevance of APP in AD and LBD, where ApoE lipoprotein is also a major genetic risk factor, investigating variations in APP levels bound to CSF lipoproteins in patients could provide insights into pathophysiological changes in these diseases. In short, it would be exceptionally interesting to study the CSF-lipoprotein proteome in AD, PD and LBD at varies stages of early and late disease. 

    It's surprising to us that in Paslawski et al. (2019), it was reported that ApoE and α-synuclein co-localized in human CSF, suggesting that α-synuclein could bind to ApoE lipoparticles. Furthermore, levels of ApoE lipoparticle-bound α-synuclein were found to be elevated in the CSF of PD patients. However, it is worth noting that the highly sensitive and high-resolution fluorescent phospholipid assay described in the current study did not detect α-synuclein in the CSF lipoprotein proteome. Investigating differences in PD patients using this approach could yield interesting findings.

    This study primarily focused on the proteome of CSF lipoproteins. The same methodology could be applied to isolate lipoproteins from CSF and analyze the lipidome by itself, potentially revealing import and significant differences within and between neurodegenerative diseases.


    . α-synuclein-lipoprotein interactions and elevated ApoE level in cerebrospinal fluid from Parkinson's disease patients. Proc Natl Acad Sci U S A. 2019 Jul 23;116(30):15226-15235. Epub 2019 Jul 3 PubMed.

  12. Aberrant lipid metabolism has been linked to several neurodegenerative diseases but, unfortunately, knowledge of the lipoprotein particles in CSF (CSF-Lps) is limited. The main obstacle is their much smaller concentration in CSF than plasma, which prohibits application of standard assays. To overcome this, Merrill et al. developed sensitive fluorescent technology to characterize CSF lipoproteins in smaller volumes and less concentrated samples. The authors went into great detail to describe how they standardized their assay using samples from plasma and comparing their profiles to their CSF samples, thus proving the utility of their methods in correctly measuring CSF lipoproteins.

    They identified that the bulk of CSF-Lps is eluting in fractions 29-30 between plasma VLDL/LDL and HDL peaks, which is consistent with a slightly higher diameter of CSF-Lps than plasma HDL. This could be explained by CSF HDL containing APOE as their major apolipoprotein, in contrast to plasma HDL, whose major apolipoprotein is APOA1.

    The authors also identified other peaks corresponding to more compact CSF-Lps with a smaller diameter, such as peak #8 (at fractions 39-40) and peak #10 (at fractions 46-47). Unfortunately, they did not perform lipidomic analysis to determine the lipid composition of each fraction, and therefore we may only speculate whether these smaller peaks represent bona fide lipoproteins.

    The authors then performed proteomics to identify which proteins are eluting in each fraction. APOE and APOJ/clusterin were associated with the main peak CSF-Lp fractions (fractions 29-30).

    Overlapping with the main peak and eluting with APOE were neuronal cell adhesion proteins NRCAM and NCAM1, and CNTN1, which also play a role in cell adhesion. These were much less abundant than APOE and APOJ, and their significance in the main fraction of CSF-Lps at present is unknown. APOA1 and APOA2 eluted slightly later than APOE, containing CSF-Lps associated with a smaller diameter CSF-Lps (eluting with fractions 35-36).

    Interestingly, the authors found no apolipoproteins in peaks #8 and #10, which again raises the question if these are genuine lipoproteins? HDL particles are generated by cholesterol efflux, and apolipoproteins such as APOA1 and APOE are the major acceptors of phospholipids and cholesterol acceptors in the formation of these lipoproteins (Rothblat et al., 2010).

    It will be interesting to characterize the smaller particles and answer the question of whether these are lipoproteins, exosomes, or other extracellular vesicles that are not lipoproteins per se. Overall, the assay developed here is a significant improvement of the methodology used to characterize CSF lipoproteins.


    . High-density lipoprotein heterogeneity and function in reverse cholesterol transport. Curr Opin Lipidol. 2010 Jun;21(3):229-38. PubMed.

Make a Comment

To make a comment you must login or register.


Mutations Citations

  1. APOE C130R (ApoE4)

Paper Citations

  1. . The HDL Proteome Watch: Compilation of studies leads to new insights on HDL function. Biochim Biophys Acta Mol Cell Biol Lipids. 2022 Feb;1867(2):159072. Epub 2021 Nov 18 PubMed.
  2. . Distinct circuit-dependent functions of presynaptic neurexin-3 at GABAergic and glutamatergic synapses. Nat Neurosci. 2015 Jul;18(7):997-1007. Epub 2015 Jun 1 PubMed.
  3. . APLP1 Is a Synaptic Cell Adhesion Molecule, Supporting Maintenance of Dendritic Spines and Basal Synaptic Transmission. J Neurosci. 2017 May 24;37(21):5345-5365. Epub 2017 Apr 27 PubMed.

External Citations

  1. CSF lipoprotein list

Further Reading

No Available Further Reading

Primary Papers

  1. . Human cerebrospinal fluid contains diverse lipoprotein subspecies enriched in proteins implicated in central nervous system health. Sci Adv. 2023 Sep;9(35):eadi5571. Epub 2023 Aug 30 PubMed.