Advertisement

Brain-derived Neurotrophic Factor in Megakaryocytes*

Open AccessPublished:March 22, 2016DOI:https://doi.org/10.1074/jbc.M116.720029
      The biosynthesis of endogenous brain-derived neurotrophic factor (BDNF) has thus far been examined in neurons where it is expressed at very low levels, in an activity-dependent fashion. In humans, BDNF has long been known to accumulate in circulating platelets, at levels far higher than in the brain. During the process of blood coagulation, BDNF is released from platelets, which has led to its extensive use as a readily accessible biomarker, under the assumption that serum levels may somehow reflect brain levels. To identify the cellular origin of BDNF in platelets, we established primary cultures of megakaryocytes, the progenitors of platelets, and we found that human and rat megakaryocytes express the BDNF gene. Surprisingly, the pattern of mRNA transcripts is similar to neurons. In the presence of thapsigargin and external calcium, the levels of the mRNA species leading to efficient BDNF translation rapidly increase. Under these conditions, pro-BDNF, the obligatory precursor of biologically active BDNF, becomes readily detectable. Megakaryocytes store BDNF in α-granules, with more than 80% of them also containing platelet factor 4. By contrast, BDNF is undetectable in mouse megakaryocytes, in line with the absence of BDNF in mouse serum. These findings suggest that alterations of BDNF levels in human serum as reported in studies dealing with depression or physical exercise may primarily reflect changes occurring in megakaryocytes and platelets, including the ability of the latter to retain and release BDNF.

      Introduction

      BDNF
      The abbreviations used are: BDNF
      brain-derived neurotrophic factor
      TPO
      thrombopoietin
      TRITC
      tetramethylrhodamine isothiocyanate
      Mk
      megakaryocyte.
      is a secretory protein regulating the development and function of neural circuits (
      • Park H.
      • Poo M.M.
      Neurotrophin regulation of neural circuit development and function.
      ,
      • Zagrebelsky M.
      • Korte M.
      Form follows function: BDNF and its involvement in sculpting the function and structure of synapses.
      ). The functional relevance of BDNF in humans is firmly established following the discovery of polymorphisms and loss-of-allele mutations associated with deficits ranging from subtle memory alterations (
      • Egan M.F.
      • Kojima M.
      • Callicott J.H.
      • Goldberg T.E.
      • Kolachana B.S.
      • Bertolino A.
      • Zaitsev E.
      • Gold B.
      • Goldman D.
      • Dean M.
      • Lu B.
      • Weinberger D.R.
      The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function.
      ) to severe symptoms early in life (
      • Gray J.
      • Yeo G.S.
      • Cox J.J.
      • Morton J.
      • Adlam A.L.
      • Keogh J.M.
      • Yanovski J.A.
      • El Gharbawy A.
      • Han J.C.
      • Tung Y.C.
      • Hodges J.R.
      • Raymond F.L.
      • O'rahilly S.
      • Farooqi I.S.
      Hyperphagia, severe obesity, impaired cognitive function, and hyperactivity associated with functional loss of one copy of the brain-derived neurotrophic factor (BDNF) gene.
      ). The cDNA sequence of BDNF predicts that like other cystine-knot proteins, BDNF is first synthesized as a precursor protein, referred to as pro-BDNF ensuring the proper formation of disulfide bridges and of a biologically active, mature neurotrophin (
      • Leibrock J.
      • Lottspeich F.
      • Hohn A.
      • Hofer M.
      • Hengerer B.
      • Masiakowski P.
      • Thoenen H.
      • Barde Y.A.
      Molecular cloning and expression of brain-derived neurotrophic factor.
      ,
      • McDonald N.Q.
      • Hendrickson W.A.
      A structural superfamily of growth factors containing a cystine knot motif.
      ). Numerous experiments with various artificial expression systems have confirmed this view, in line with the results of early experiments with nerve growth factor (
      • Suter U.
      • Heymach Jr., J.V.
      • Shooter E.M.
      Two conserved domains in the NGF propeptide are necessary and sufficient for the biosynthesis of correctly processed and biologically active NGF.
      ). So far, only a very small number of studies have been performed addressing the question of the biosynthesis, storage, and secretion of endogenous BDNF (
      • Matsumoto T.
      • Rauskolb S.
      • Polack M.
      • Klose J.
      • Kolbeck R.
      • Korte M.
      • Barde Y.A.
      Biosynthesis and processing of endogenous BDNF: CNS neurons store and secrete BDNF, not pro-BDNF.
      ,
      • Yang J.
      • Siao C.J.
      • Nagappan G.
      • Marinic T.
      • Jing D.
      • McGrath K.
      • Chen Z.Y.
      • Mark W.
      • Tessarollo L.
      • Lee F.S.
      • Lu B.
      • Hempstead B.L.
      Neuronal release of proBDNF.
      ). As a result of the scarcity of the protein in neurons, most studies used instead overexpression paradigms, leading to uncertainties as to whether the processing of pro-BDNF takes place within neurons or also in the extracellular space following the secretion of pro-BDNF. Human platelets contain between 100–1,000-fold more BDNF than brain tissue when brain and platelets are compared on a protein basis (
      • Yamamoto H.
      • Gurney M.E.
      Human platelets contain brain-derived neurotrophic factor.
      ,
      • Barde Y.A.
      • Edgar D.
      • Thoenen H.
      Purification of a new neurotrophic factor from mammalian brain.
      ,
      • Burnouf T.
      • Kuo Y.P.
      • Blum D.
      • Burnouf S.
      • Su C.Y.
      Human platelet concentrates: a source of solvent/detergent-treated highly enriched brain-derived neurotrophic factor.
      ). As it appears unlikely that the biosynthesis of BDNF takes place in platelets, we established primary cultures of megakaryocytes (Mks), the progenitors of platelets. Beyond questions related to the biosynthesis of endogenous BDNF and to the productive expression of its gene in non-neuronal cells, the question of the origin of BDNF in human blood and serum is of wider interest. Indeed, BDNF levels in human serum are widely used as a biomarker speculated to somehow reflect brain levels. Thus, countless studies have reported decreased BDNF levels in serum in mood disorders, including depression (
      • Munkholm K.
      • Vinberg M.
      • Kessing L.V.
      Peripheral blood brain-derived neurotrophic factor in bipolar disorder: a comprehensive systematic review and meta-analysis.
      ), although by contrast physical exercise has been found to increase them (
      • Szuhany K.L.
      • Bugatti M.
      • Otto M.W.
      A meta-analytic review of the effects of exercise on brain-derived neurotrophic factor.
      ).

      Results

      To determine whether BDNF can be detected in Mks, we established primary cultures of mouse, rat, and human Mks and analyzed their content by Western blotting using a BDNF monoclonal antibody specifically recognizing BDNF and not NGF or NT4, under conditions where both are recognized by their corresponding antibodies (data not shown). A strong signal corresponding to the expected size of the monomer of mature BDNF was detected in both rat and human but not in mouse Mks (Fig. 1A). Note that pro-BDNF is barely detectable under these conditions, with only a faint band detected in human Mks. As a control, we also analyzed the BDNF content of platelets purified from the corresponding species and confirmed the presence of considerable amounts of BDNF in human platelets, with substantial levels also detected in rat but not in mouse platelets (Fig. 1B). This result with platelets is in agreement with previous conclusions using immunoassay determinations with the serum prepared from these three species (
      • Radka S.F.
      • Holst P.A.
      • Fritsche M.
      • Altar C.A.
      Presence of brain-derived neurotrophic factor in brain and human and rat but not mouse serum detected by a sensitive and specific immunoassay.
      ). We then explored whether BDNF would be localized in α-granules, the storage compartment of a number of growth factors in the Mk lineage. A BDNF-specific signal was found to extensively co-localize with PF4, also designated CXCl4, one of the most abundant α-granule proteins (Fig. 1C). A distinct BDNF signal could also be seen in the tips of the proplatelet-forming human Mks (Fig. 1D, arrows), consistent with the notion that BDNF is transferred from Mks to platelets. As a control for the specificity of the BDNF signal in these immunostaining experiments, we used mouse Mks. When incubated with the corresponding antibodies under the same experimental conditions as rat and human Mks, PF4, but not BDNF, was detected (Fig. 1C). The lack of any detectable BDNF in mouse platelets is not a feature specific to CD1 mouse strain that we used in most of our experiments, as the C57BL/6 strain led to identically negative results, both in Mks and platelets. We note that in a recent highly sensitive and quantitative proteomic analysis of C57BL/6 platelet extracts, no BDNF could be detected (
      • Zeiler M.
      • Moser M.
      • Mann M.
      Copy number analysis of the murine platelet proteome spanning the complete abundance range.
      ). Quantification of the BDNF levels in rat Mks by ELISA indicated that their lysates contain 1.40 ± 0.13 ng/mg protein (mean ± S.E., n = 6). We also attempted to determine whether proplatelet-forming Mks release BDNF into the medium by incubating the cells with HRP-conjugated BDNF monoclonal antibodies (see “Experimental Procedures”). After 2 days of incubation with the BDNF capture antibody, we failed to detect any release of BDNF into the Mk-conditioned medium, suggesting that the bulk of BDNF is transferred into platelets and not released into the medium. This result is consistent with the previous work indicating that both in rats and humans the levels of BDNF are far higher in serum than in plasma, suggesting that the bulk of BDNF in serum results from platelet degranulation (
      • Radka S.F.
      • Holst P.A.
      • Fritsche M.
      • Altar C.A.
      Presence of brain-derived neurotrophic factor in brain and human and rat but not mouse serum detected by a sensitive and specific immunoassay.
      ). To determine whether the BDNF biosynthetic machinery is expressed in Mks, we isolated RNA from mature platelet-forming mouse, rat, and human Mks. Significant levels of BDNF mRNAs were detected in both rat and in human cells (Fig. 2A). In four separate experiments, the total mRNA levels in both species were found to be about 200-fold higher in rat and human, compared with mouse Mks. All known BDNF transcripts present in RNA extracted from Mks of the three species were analyzed and compared with RNA extracted from the hippocampus and the lung as neuronal and non-neuronal reference tissues of the corresponding species (Fig. 2A). The results of these experiments revealed a neuronal pattern of mRNA expression in both rat and human Mks, with a prominent inclusion of exon I and IV transcripts. Notably, exon I-containing transcripts have recently been shown to markedly increase Bdnf mRNA translatability (
      • Koppel I.
      • Tuvikene J.
      • Lekk I.
      • Timmusk T.
      Efficient use of a translation start codon in BDNF exon I.
      ). By contrast, these typical neuronal transcripts were undetectable in the mouse (Fig. 2A). The levels of all main transcripts were also assessed in the three species by real time PCR (Fig. 2B).
      Figure thumbnail gr1
      Figure 1Differential BDNF protein levels in mouse, rat, and human megakaryocytes and platelets. Western blot lysates of cultured Mks (A) and blood platelets (B) are shown. Eighty micrograms of protein per lane were loaded, and the blotting membrane was incubated with the mouse monoclonal antibody 3C11 developed by Icosagen (Tartu, Estonia). Recombinant BDNF and pro-BDNF were used as molecular mass markers and antibodies to β-actin as loading controls. Asterisks (top right panels) point to a band unrelated to BDNF likely corresponding to immunoglobulin light chains in the mouse sample. Note the absence of BDNF in mouse Mks and platelets. C, antibodies to BDNF 9 (green) (
      • Dieni S.
      • Matsumoto T.
      • Dekkers M.
      • Rauskolb S.
      • Ionescu M.S.
      • Deogracias R.
      • Gundelfinger E.D.
      • Kojima M.
      • Nestel S.
      • Frotscher M.
      • Barde Y.A.
      BDNF and its pro-peptide are stored in presynaptic dense core vesicles in brain neurons.
      ) and PF4 (red) reveal expression of both antigens in mature rat and human Mks. Note that unlike PF4, BDNF is not detectable in mouse Mks. The co-localization of BDNF with PF4 in rat and human Mks was quantified using the pixel intensity specifically generated by each channel. In humans, 83% and in rats 86% BDNF-positive granules were also PF4-positive. Blue, DAPI staining. D, immunofluorescence staining of F-actin (red) and BDNF (green) in proplatelet-forming cultured human Mks. Arrows indicate BDNF accumulation in proplatelet buds.
      Figure thumbnail gr2
      Figure 2Transcriptional analysis of BDNF in mouse, rat, and human megakaryocytes. Conventional (A) and real time quantitative (B) PCR using exon-specific primers with RNA extracted from mature cultured Mks, adult hippocampus (Hippo), and lung are shown. Note that in the mouse, the neuron-specific transcripts, including exon I and IV, are not detected and that by contrast the transcript pattern resembles the non-neuronal pattern observed in lung tissue. The converse is the case with RNA extracted from rat and human Mks with transcript patterns, including exon I and IV, that are characteristic of a neuronal pattern as illustrated with the hippocampus. Unless indicated as non-significant (n.s.), all values are mean values ± S.E. in triplicates and based on three independent experiments, at p < 0.001 (paired t test).
      In neurons, increased levels of cytoplasmic calcium have long been known to activate Bdnf transcription by activating promoters I and IV in particular (
      • West A.E.
      • Pruunsild P.
      • Timmusk T.
      Neurotrophins: transcription and translation.
      ). As Mks are devoid of voltage-activated calcium channels, the addition of thapsigargin to non-electrically excitable cells such as Mks offered an opportunity to test whether the corresponding Bdnf promoters are also responsive to calcium levels in Mks. Thapsigargin is a selective inhibitor of the sarco/endoplasmic reticulum Ca2+-ATPase, secondarily leading to the opening of stored-operated calcium channels at the cell surface and increased levels of cytoplasmic calcium (
      • Cheng K.T.
      • Ong H.L.
      • Liu X.
      • Ambudkar I.S.
      Contribution of TRPC1 and Orai1 to Ca2+ entry activated by store depletion.
      ). We found that at nanomolar concentrations (Fig. 3A), thapsigargin massively activates transcription in a time-dependent manner (Fig. 3B). This increase was primarily accounted for by contributions from exons I, IV, and IXa and to a lesser extent exon VI (Fig. 3C). We then tested whether complexing extracellular calcium would decrease the thapsigargin-induced transcriptional activation of Bdnf and found that 2.5 mm EGTA completely blocked the inductive effects of thapsigargin as assessed both by primers corresponding to the protein coding sequence or by exon-specific primers (Fig. 3C). To further test the notion that intracellular calcium levels regulate Bdnf transcription, we tested the effects of the calcium ionophore ionomycin. When added to mature Mk cultures for 4 h, it increased the levels of Bdnf mRNA by 10.50 ± 1.85-fold (n = 3, mean ± S.E.). As still very little is known about the biosynthesis of endogenous BDNF in any cell type, we were then curious to see how the Mk translation and processing machinery would cope with the massive increase in Bdnf mRNA levels caused by thapsigargin. We found that at 10 nm, thapsigargin led not only to a marked increase of processed (or mature) BDNF but also to readily detectable levels of pro-BDNF, suggesting that the thapsigargin-induced increased transcription may temporarily saturate the pro-BDNF processing capacity of Mks (Fig. 4, A–C). The 3C11 BDNF monoclonal antibody recognizes not only mature but also (as expected) unprocessed and partially processed forms of BDNF as indicated in Fig. 4, A and B. In previous experiments using heterologous expression systems, we noted that the replacement of an arginine residue in position −1 by lysine at the furin cleavage site of pro-BDNF led to the accumulation of an N-glycosylated intermediate product corresponding to the size indicated by the arrows in Fig. 4, A and B. Amino-terminal sequencing of this product indicated that the use of an alternative cleavage site generated a product with a 15-residue addition to the amino terminus of mature BDNF (
      • Kolbeck R.
      • Jungbluth S.
      • Barde Y.A.
      Characterisation of neurotrophin dimers and monomers.
      ). This product was shown to be N-glycosylated (
      • West A.E.
      • Pruunsild P.
      • Timmusk T.
      Neurotrophins: transcription and translation.
      ). The use of the pro-BDNF monoclonal antibody H1001G independently confirmed the identity of pro-BDNF and also allowed the detection of the BDNF cleaved pro-peptide (Fig. 4C). We note that in the absence of thapsigargin stimulation, the steady state levels of pro-BDNF are even lower than in neurons where we estimated them to represent about 1 molecule for 10 molecules of mature BDNF (
      • Matsumoto T.
      • Rauskolb S.
      • Polack M.
      • Klose J.
      • Kolbeck R.
      • Korte M.
      • Barde Y.A.
      Biosynthesis and processing of endogenous BDNF: CNS neurons store and secrete BDNF, not pro-BDNF.
      ,
      • Dieni S.
      • Matsumoto T.
      • Dekkers M.
      • Rauskolb S.
      • Ionescu M.S.
      • Deogracias R.
      • Gundelfinger E.D.
      • Kojima M.
      • Nestel S.
      • Frotscher M.
      • Barde Y.A.
      BDNF and its pro-peptide are stored in presynaptic dense core vesicles in brain neurons.
      ). We also tested the expression of the most likely protease candidate thought to be involved in the generation of mature BDNF, namely furin and the proprotein convertase 7, the latter having been recently shown to be necessary for the processing of BDNF in neurons (
      • Wetsel W.C.
      • Rodriguiz R.M.
      • Guillemot J.
      • Rousselet E.
      • Essalmani R.
      • Kim I.H.
      • Bryant J.C.
      • Marcinkiewicz J.
      • Desjardins R.
      • Day R.
      • Constam D.B.
      • Prat A.
      • Seidah N.G.
      Disruption of the expression of the proprotein convertase PC7 reduces BDNF production and affects learning and memory in mice.
      ). The results of conventional (Fig. 5A) and real time quantitative PCR (Fig. 5B) experiments indicated that both enzymes are present in rat Mks, at levels comparable with those found in the rat hippocampus (Fig. 5). In line with this, the processing of pro-BDNF was largely prevented by the addition of the convertase inhibitor decanoyl-Arg-Val-Lys-Arg-chloromethyl ketone (data not shown).
      Figure thumbnail gr3
      Figure 3Up-regulation of Bdnf mRNA by thapsigargin. Effect of extracellular calcium. Dose response (A) and time course (B) of Bdnf mRNA expression by rat Mks after thapsigargin treatment. Purified mature rat Mks were cultured in the presence or absence of thapsigargin or vehicle (DMSO) used at the indicated concentrations (A) and for different lengths of times (B). Total mRNA was extracted and reverse-transcribed, and the resulting cDNA was amplified by real time quantitative PCR using specific primers for the coding sequence of Bdnf. C, extracellular calcium dependence of thapsigargin-induced Bdnf mRNA increase. Rat Mks were preincubated with 2.5 mm EGTA for 1.5 h at 37 °C followed by 10 nm thapsigargin for 4 h. mRNA expression was analyzed by real time quantitative PCR using specific primers for the coding sequence of Bdnf (CDS) or exon-specific primers. All values are mean values ± S.E. in triplicates and based on three independent experiments. Unless indicated, all the statistical values are compared with the control. *, p < 0.05; ***, p < 0.001 (paired t test).
      Figure thumbnail gr4
      Figure 4Effect of thapsigargin on pro-BDNF, mature BDNF, and pro-peptide in rat Mks. Dose response (A) and time course (B) of pro-BDNF and mature BDNF proteins by rat Mks after thapsigargin treatment are shown. Mature Mks were cultured for 16 h at the indicated doses of thapsigargin (A) or 10 nm thapsigargin for the indicated times (B). Forty micrograms of protein per lane were loaded, and the blotting membrane was incubated with the mouse monoclonal antibody 3C11 developed by Icosagen. Arrows indicate intermediate proteolytic products of pro-BDNF (C). Time course of pro-BDNF and pro-peptide proteins generated by rat Mks incubated with 10 nm thapsigargin for the indicated time periods. Eighty micrograms protein per lane were loaded, and the blotting membrane was incubated with the mouse monoclonal antibody H1001G developed by GeneCopeia, Inc. The blots shown are representative of three independent experiments with similar results. Graphs show mean ± S.E. of the densitometric values quantified from the blots of the three separate experiments. ***, p < 0.001 (paired t test compared the corresponding controls). Recombinant BDNF (150–300 pg), cleavage-resistant recombinant pro-BDNF (0.5–1 ng), and recombinant pro-peptide (1–10 ng) were used as molecular mass markers and antibodies to β-actin as loading controls.
      Figure thumbnail gr5
      Figure 5Differential expression of proprotein convertases in primary MKs. Conventional (A) and real time quantitative (B) PCRs using specific primers and RNA extracted from mature cultured rat Mks and adult hippocampus are shown. Note that although transcripts, including Pcsk1, PcsK2, Pcsk4, and Pcsk5, are expressed in hippocampal tissue, they are not detected in Mks (A). Comparative expression levels between the two tissues for the expressed proprotein convertases are shown in B. All values are mean values ± S.E. in triplicates and are based on three independent experiments.

      Discussion

      The results obtained with primary cultures of Mks suggest that these cells represent the main source of BDNF in platelets as well as in serum. First, Mks contain readily detectable levels of BDNF protein. Second, BDNF is stored in α-granules in Mks, long known to also represent the storage compartment of various growth factors and cytokines in platelets (
      • Maynard D.M.
      • Heijnen H.F.
      • Horne M.K.
      • White J.G.
      • Gahl W.A.
      Proteomic analysis of platelet α-granules using mass spectrometry.
      ). Third, BDNF can be visualized in proplatelets (Fig. 1D) suggesting that platelets contain BDNF by the time they begin to separate from Mks. Fourth, the BDNF gene is expressed at relatively high levels in rat and human Mks, although by contrast, the levels are much lower in mice. Notably, the mouse transcripts do not include exon I, which allows efficient translation of BDNF mRNA (
      • Koppel I.
      • Tuvikene J.
      • Lekk I.
      • Timmusk T.
      Efficient use of a translation start codon in BDNF exon I.
      ). This negative result with mouse Mks is significant as it correlates with the lack of detectable levels of BDNF in mouse serum, unlike the case with rat and human sera (
      • Radka S.F.
      • Holst P.A.
      • Fritsche M.
      • Altar C.A.
      Presence of brain-derived neurotrophic factor in brain and human and rat but not mouse serum detected by a sensitive and specific immunoassay.
      ). Taken together, it would seem that circulating platelets, once filled up with BDNF packaged in α-granules inherited from Mks, represent the only significant source of BDNF in serum. Other sources such as endothelial or immune cells that do express the BDNF gene at low levels (
      • Kerschensteiner M.
      • Gallmeier E.
      • Behrens L.
      • Leal V.V.
      • Misgeld T.
      • Klinkert W.E.
      • Kolbeck R.
      • Hoppe E.
      • Oropeza-Wekerle R.L.
      • Bartke I.
      • Stadelmann C.
      • Lassmann H.
      • Wekerle H.
      • Hohlfeld R.
      Activated human T cells, B cells, and monocytes produce brain-derived neurotrophic factor in vitro and in inflammatory brain lesions: a neuroprotective role of inflammation?.
      ,
      • Snapyan M.
      • Lemasson M.
      • Brill M.S.
      • Blais M.
      • Massouh M.
      • Ninkovic J.
      • Gravel C.
      • Berthod F.
      • Götz M.
      • Barker P.A.
      • Parent A.
      • Saghatelyan A.
      Vasculature guides migrating neuronal precursors in the adult mammalian forebrain via brain-derived neurotrophic factor signaling.
      ) do not seem to make significant contributions to circulating levels of BDNF as the levels are undetectable in mouse serum (
      • Radka S.F.
      • Holst P.A.
      • Fritsche M.
      • Altar C.A.
      Presence of brain-derived neurotrophic factor in brain and human and rat but not mouse serum detected by a sensitive and specific immunoassay.
      ). Also, the very low levels of BDNF found in human and rat plasma may in fact be accounted for by microparticles or exosomes released from platelets (
      • Heijnen H.F.
      • Schiel A.E.
      • Fijnheer R.
      • Geuze H.J.
      • Sixma J.J.
      Activated platelets release two types of membrane vesicles: microvesicles by surface shedding and exosomes derived from exocytosis of multivesicular bodies and α-granules.
      ). It is intriguing to note that human platelets and sera contain BDNF levels that are about 10 times higher than in the rat. The reasons for this difference are unclear at this point, as is the function of BDNF in platelets. The lack of BDNF in mouse platelets suggests that whatever the biological role of platelet-derived BDNF may be, it could be redundant in the mouse with the function of other platelet-derived growth factors. A signaling system based on the circulation of small and ubiquitous cellular fragments, including exosomes loaded up with a powerful neurotrophic factor, has the potential to be functionally relevant in the context of human brain function. As blood flow is tightly regulated by neuronal activity (
      • Hillman E.M.
      Coupling mechanism and significance of the BOLD signal: a status report.
      ), it is conceivable that in humans exosomes loaded with BDNF (
      • Aatonen M.T.
      • Ohman T.
      • Nyman T.A.
      • Laitinen S.
      • Gronholm M.
      • Siljander P.R.
      Isolation and characterization of platelet-derived extracellular vesicles.
      ) may be delivered to the brain in activity-dependent fashion, possibly explaining the beneficial effects of physical exercise. Although this is a matter of speculation at this point, the possibility of a functional role for BDNF in platelets can now be tested by engineering the mouse genome so as to replicate the situation in humans. Alternatively, it is also conceivable that the functional significance of BDNF in human platelets may remain as mysterious as that of NGF and EGF in the adult male mouse submandibular gland (
      • Cohen S.
      Origins of growth factors: NGF and EGF.
      ).
      It may seem surprising that the cellular source of BDNF in rat and human serum has not been previously uncovered, especially in view of the very extensive use of BDNF as a biomarker in human blood (
      • Polyakova M.
      • Stuke K.
      • Schuemberg K.
      • Mueller K.
      • Schoenknecht P.
      • Schroeter M.L.
      BDNF as a biomarker for successful treatment of mood disorders: a systematic & quantitative meta-analysis.
      ). An analysis of the corresponding literature reveals that negative results were obtained early on in experiments specifically addressing the question of BDNF expression in human Mks (
      • Kolbeck R.
      • Bartke I.
      • Eberle W.
      • Barde Y.A.
      Brain-derived neurotrophic factor levels in the nervous system of wild-type and neurotrophin gene mutant mice.
      ). These experiments were performed with the megakaryocyte lines DAMI and Meg-01 and led to the conclusion that the BDNF gene is not expressed in the cells (
      • Kolbeck R.
      • Bartke I.
      • Eberle W.
      • Barde Y.A.
      Brain-derived neurotrophic factor levels in the nervous system of wild-type and neurotrophin gene mutant mice.
      ). Although we confirmed these results, it appears plausible that these tumor lines fail to faithfully replicate late aspects of Mk maturation, as is not rarely the case with readily expandable tumor cells. Following this negative result, the presence of BDNF in platelets has been speculated to results from a hypothetical uptake from sources such as the brain. However, this notion has not been substantiated by plausible mechanisms, unlike in the case of serotonin, a neurotransmitter long known to accumulate in the dense granules of platelets following its uptake by specific transporters located in the membrane of platelets.
      The identification of Mks as the source of BDNF in platelets invites a revision of the widely held view that in humans the serum levels of BDNF reflect its levels in the brain. In addition to our findings, it has long been established that radiolabeled BDNF does not reach the brain when injected into the peripheral circulation (
      • Pardridge W.M.
      • Kang Y.S.
      • Buciak J.L.
      Transport of human recombinant brain-derived neurotrophic factor (BDNF) through the rat blood-brain barrier in vivo using vector-mediated peptide drug delivery.
      ). It appears then that the variations in the levels of BDNF reported in various conditions, including the increase after physical exercise (
      • Szuhany K.L.
      • Bugatti M.
      • Otto M.W.
      A meta-analytic review of the effects of exercise on brain-derived neurotrophic factor.
      ) or decrease during the course of depressive episodes (
      • Munkholm K.
      • Vinberg M.
      • Kessing L.V.
      Peripheral blood brain-derived neurotrophic factor in bipolar disorder: a comprehensive systematic review and meta-analysis.
      ), are in need of alternative plausible explanations, and it is conceivable that these variations may reflect different degrees of platelet activation (
      • Kestin A.S.
      • Ellis P.A.
      • Barnard M.R.
      • Errichetti A.
      • Rosner B.A.
      • Michelson A.D.
      Effect of strenuous exercise on platelet activation state and reactivity.
      ). In addition, there is emerging evidence that the hematopoietic niche where Mks develop (
      • Day R.B.
      • Link D.C.
      Megakaryocytes in the hematopoietic stem cell niche.
      ) is innervated by the peripheral nervous system and that hematopoietic cells may respond to nerve-derived signals (
      • Méndez-Ferrer S.
      • Michurina T.V.
      • Ferraro F.
      • Mazloom A.R.
      • Macarthur B.D.
      • Lira S.A.
      • Scadden D.T.
      • Ma'ayan A.
      • Enikolopov G.N.
      • Frenette P.S.
      Mesenchymal and haematopoietic stem cells form a unique bone marrow niche.
      ). However, whether or not these stimuli change the expression levels of BDNF in Mks remains unclear at this point.
      With regard to the biosynthesis of endogenous BDNF, our results suggest that Mks could represent an alternative cellular model to neurons, which have been so difficult to study in the face of the very low levels of expression of BDNF levels in these cells. By comparison with BDNF levels in the brain (
      • Barde Y.A.
      • Edgar D.
      • Thoenen H.
      Purification of a new neurotrophic factor from mammalian brain.
      ), the levels of BDNF in human platelets are significantly higher, 100–1,000-fold on a per mg of protein basis when brain extracts and purified platelets are compared (
      • Burnouf T.
      • Kuo Y.P.
      • Blum D.
      • Burnouf S.
      • Su C.Y.
      Human platelet concentrates: a source of solvent/detergent-treated highly enriched brain-derived neurotrophic factor.
      ). In particular, Mks offer a new opportunity to examine the biosynthesis of endogenous BDNF and the possible role of pro-BDNF. Transcription activation by thapsigargin leads to clearly detectable levels of pro-BDNF without the need for prior enrichment by immunoprecipitation. In view of the current interest related to the Val → Met substitution in pro-BDNF (
      • Egan M.F.
      • Kojima M.
      • Callicott J.H.
      • Goldberg T.E.
      • Kolachana B.S.
      • Bertolino A.
      • Zaitsev E.
      • Gold B.
      • Goldman D.
      • Dean M.
      • Lu B.
      • Weinberger D.R.
      The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function.
      ), human Mks of the corresponding genotype may represent an interesting cellular model to understand the biochemical consequences of this amino acid replacement.
      In conclusion, our results contribute to clarify the cellular origin of BDNF in human blood; and they describe a tractable cellular system to study the biosynthesis of endogenous BDNF.

      Author Contributions

      P. C. F. designed, performed, and analyzed the experiments illustrated in Figure 1, Figure 2, Figure 3, Figure 4, Figure 5 and K. S. was also involved. T. M., M. C., and C. G. were involved in all aspects of the work related to human, rat, and mouse Mks and in the interpretation of the results. Y. A. B. helped with the initiation of the project, the design of the experiments, and with the interpretation of the results. All have read the manuscript and discussed its content. The final version was approved by all.

      Acknowledgments

      We thank Mike Greenberg, Michael Frotscher, and Hayley Dingsdale for useful discussions.

      Author Profile

      References

        • Park H.
        • Poo M.M.
        Neurotrophin regulation of neural circuit development and function.
        Nat. Rev. Neurosci. 2013; 14: 7-23
        • Zagrebelsky M.
        • Korte M.
        Form follows function: BDNF and its involvement in sculpting the function and structure of synapses.
        Neuropharmacology. 2014; 76: 628-638
        • Egan M.F.
        • Kojima M.
        • Callicott J.H.
        • Goldberg T.E.
        • Kolachana B.S.
        • Bertolino A.
        • Zaitsev E.
        • Gold B.
        • Goldman D.
        • Dean M.
        • Lu B.
        • Weinberger D.R.
        The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function.
        Cell. 2003; 112: 257-269
        • Gray J.
        • Yeo G.S.
        • Cox J.J.
        • Morton J.
        • Adlam A.L.
        • Keogh J.M.
        • Yanovski J.A.
        • El Gharbawy A.
        • Han J.C.
        • Tung Y.C.
        • Hodges J.R.
        • Raymond F.L.
        • O'rahilly S.
        • Farooqi I.S.
        Hyperphagia, severe obesity, impaired cognitive function, and hyperactivity associated with functional loss of one copy of the brain-derived neurotrophic factor (BDNF) gene.
        Diabetes. 2006; 55: 3366-3371
        • Leibrock J.
        • Lottspeich F.
        • Hohn A.
        • Hofer M.
        • Hengerer B.
        • Masiakowski P.
        • Thoenen H.
        • Barde Y.A.
        Molecular cloning and expression of brain-derived neurotrophic factor.
        Nature. 1989; 341: 149-152
        • McDonald N.Q.
        • Hendrickson W.A.
        A structural superfamily of growth factors containing a cystine knot motif.
        Cell. 1993; 73: 421-424
        • Suter U.
        • Heymach Jr., J.V.
        • Shooter E.M.
        Two conserved domains in the NGF propeptide are necessary and sufficient for the biosynthesis of correctly processed and biologically active NGF.
        EMBO J. 1991; 10: 2395-2400
        • Matsumoto T.
        • Rauskolb S.
        • Polack M.
        • Klose J.
        • Kolbeck R.
        • Korte M.
        • Barde Y.A.
        Biosynthesis and processing of endogenous BDNF: CNS neurons store and secrete BDNF, not pro-BDNF.
        Nat. Neurosci. 2008; 11: 131-133
        • Yang J.
        • Siao C.J.
        • Nagappan G.
        • Marinic T.
        • Jing D.
        • McGrath K.
        • Chen Z.Y.
        • Mark W.
        • Tessarollo L.
        • Lee F.S.
        • Lu B.
        • Hempstead B.L.
        Neuronal release of proBDNF.
        Nat. Neurosci. 2009; 12: 113-115
        • Yamamoto H.
        • Gurney M.E.
        Human platelets contain brain-derived neurotrophic factor.
        J. Neurosci. 1990; 10: 3469-3478
        • Barde Y.A.
        • Edgar D.
        • Thoenen H.
        Purification of a new neurotrophic factor from mammalian brain.
        EMBO J. 1982; 1: 549-553
        • Burnouf T.
        • Kuo Y.P.
        • Blum D.
        • Burnouf S.
        • Su C.Y.
        Human platelet concentrates: a source of solvent/detergent-treated highly enriched brain-derived neurotrophic factor.
        Transfusion. 2012; 52: 1721-1728
        • Munkholm K.
        • Vinberg M.
        • Kessing L.V.
        Peripheral blood brain-derived neurotrophic factor in bipolar disorder: a comprehensive systematic review and meta-analysis.
        Mol. Psychiatry. 2016; 21: 216-228
        • Szuhany K.L.
        • Bugatti M.
        • Otto M.W.
        A meta-analytic review of the effects of exercise on brain-derived neurotrophic factor.
        J. Psychiatr. Res. 2015; 60: 56-64
        • Dieni S.
        • Matsumoto T.
        • Dekkers M.
        • Rauskolb S.
        • Ionescu M.S.
        • Deogracias R.
        • Gundelfinger E.D.
        • Kojima M.
        • Nestel S.
        • Frotscher M.
        • Barde Y.A.
        BDNF and its pro-peptide are stored in presynaptic dense core vesicles in brain neurons.
        J. Cell Biol. 2012; 196: 775-788
        • Kolbeck R.
        • Bartke I.
        • Eberle W.
        • Barde Y.A.
        Brain-derived neurotrophic factor levels in the nervous system of wild-type and neurotrophin gene mutant mice.
        J. Neurochem. 1999; 72: 1930-1938
        • Kwok S.C.
        • Chakraborty D.
        • Soares M.J.
        • Dai G.
        Relative expression of proprotein convertases in rat ovaries during pregnancy.
        J. Ovarian Res. 2013; 6: 91
        • Radka S.F.
        • Holst P.A.
        • Fritsche M.
        • Altar C.A.
        Presence of brain-derived neurotrophic factor in brain and human and rat but not mouse serum detected by a sensitive and specific immunoassay.
        Brain Res. 1996; 709: 122-301
        • Zeiler M.
        • Moser M.
        • Mann M.
        Copy number analysis of the murine platelet proteome spanning the complete abundance range.
        Mol. Cell. Proteomics. 2014; 13: 3435-3445
        • Koppel I.
        • Tuvikene J.
        • Lekk I.
        • Timmusk T.
        Efficient use of a translation start codon in BDNF exon I.
        J. Neurochem. 2015; 134: 1015-1025
        • West A.E.
        • Pruunsild P.
        • Timmusk T.
        Neurotrophins: transcription and translation.
        Handb. Exp. Pharmacol. 2014; 220: 67-100
        • Cheng K.T.
        • Ong H.L.
        • Liu X.
        • Ambudkar I.S.
        Contribution of TRPC1 and Orai1 to Ca2+ entry activated by store depletion.
        Adv. Exp. Med. Biol. 2011; 704: 435-449
        • Kolbeck R.
        • Jungbluth S.
        • Barde Y.A.
        Characterisation of neurotrophin dimers and monomers.
        Eur. J. Biochem. 1994; 225: 995-1003
        • Wetsel W.C.
        • Rodriguiz R.M.
        • Guillemot J.
        • Rousselet E.
        • Essalmani R.
        • Kim I.H.
        • Bryant J.C.
        • Marcinkiewicz J.
        • Desjardins R.
        • Day R.
        • Constam D.B.
        • Prat A.
        • Seidah N.G.
        Disruption of the expression of the proprotein convertase PC7 reduces BDNF production and affects learning and memory in mice.
        Proc. Natl. Acad. Sci. U.S.A. 2013; 110: 17362-17367
        • Maynard D.M.
        • Heijnen H.F.
        • Horne M.K.
        • White J.G.
        • Gahl W.A.
        Proteomic analysis of platelet α-granules using mass spectrometry.
        J. Thromb. Haemost. 2007; 5: 1945-1955
        • Kerschensteiner M.
        • Gallmeier E.
        • Behrens L.
        • Leal V.V.
        • Misgeld T.
        • Klinkert W.E.
        • Kolbeck R.
        • Hoppe E.
        • Oropeza-Wekerle R.L.
        • Bartke I.
        • Stadelmann C.
        • Lassmann H.
        • Wekerle H.
        • Hohlfeld R.
        Activated human T cells, B cells, and monocytes produce brain-derived neurotrophic factor in vitro and in inflammatory brain lesions: a neuroprotective role of inflammation?.
        J. Exp. Med. 1999; 189: 865-870
        • Snapyan M.
        • Lemasson M.
        • Brill M.S.
        • Blais M.
        • Massouh M.
        • Ninkovic J.
        • Gravel C.
        • Berthod F.
        • Götz M.
        • Barker P.A.
        • Parent A.
        • Saghatelyan A.
        Vasculature guides migrating neuronal precursors in the adult mammalian forebrain via brain-derived neurotrophic factor signaling.
        J. Neurosci. 2009; 29: 4172-4188
        • Heijnen H.F.
        • Schiel A.E.
        • Fijnheer R.
        • Geuze H.J.
        • Sixma J.J.
        Activated platelets release two types of membrane vesicles: microvesicles by surface shedding and exosomes derived from exocytosis of multivesicular bodies and α-granules.
        Blood. 1999; 94: 3791-3799
        • Hillman E.M.
        Coupling mechanism and significance of the BOLD signal: a status report.
        Annu. Rev. Neurosci. 2014; 37: 161-181
        • Aatonen M.T.
        • Ohman T.
        • Nyman T.A.
        • Laitinen S.
        • Gronholm M.
        • Siljander P.R.
        Isolation and characterization of platelet-derived extracellular vesicles.
        J. Extracell. Vesicles. 2014; 3https://doi.org/10.3402/jev.v3.24692
        • Cohen S.
        Origins of growth factors: NGF and EGF.
        J. Biol. Chem. 2008; 283: 33793-33797
        • Polyakova M.
        • Stuke K.
        • Schuemberg K.
        • Mueller K.
        • Schoenknecht P.
        • Schroeter M.L.
        BDNF as a biomarker for successful treatment of mood disorders: a systematic & quantitative meta-analysis.
        J. Affect. Disord. 2015; 174: 432-440
        • Pardridge W.M.
        • Kang Y.S.
        • Buciak J.L.
        Transport of human recombinant brain-derived neurotrophic factor (BDNF) through the rat blood-brain barrier in vivo using vector-mediated peptide drug delivery.
        Pharm. Res. 1994; 11: 738-746
        • Kestin A.S.
        • Ellis P.A.
        • Barnard M.R.
        • Errichetti A.
        • Rosner B.A.
        • Michelson A.D.
        Effect of strenuous exercise on platelet activation state and reactivity.
        Circulation. 1993; 88: 1502-1511
        • Day R.B.
        • Link D.C.
        Megakaryocytes in the hematopoietic stem cell niche.
        Nat. Med. 2014; 20: 1233-1234
        • Méndez-Ferrer S.
        • Michurina T.V.
        • Ferraro F.
        • Mazloom A.R.
        • Macarthur B.D.
        • Lira S.A.
        • Scadden D.T.
        • Ma'ayan A.
        • Enikolopov G.N.
        • Frenette P.S.
        Mesenchymal and haematopoietic stem cells form a unique bone marrow niche.
        Nature. 2010; 466: 829-834