ask1 inhibits astroglial development via p38 mitogen-activated

MOLECULAR AND CELLULAR BIOLOGY, Jan. 2004, p. 280–293 Vol. 24, No. 10270-7306/04/$08.00�0 DOI: 10.1128/MCB.24.1.280–293.2004Copyright © 2004, American Society for Microbiology. All Rights Reserved.

ASK1 Inhibits Astroglial Development via p38 Mitogen-ActivatedProtein Kinase and Promotes Neuronal Differentiation in Adult

Hippocampus-Derived Progenitor CellsRoland Faigle, Anke Brederlau, Muna Elmi, Yvonne Arvidsson, Tatsuo S. Hamazaki,†

Hidetaka Uramoto, and Keiko Funa*Department of Medical Cell Biology, Institute of Anatomy and Cell Biology, Goteborg University,

Gothenburg, Sweden

Received 24 March 2003/Returned for modification 25 June 2003/Accepted 1 October 2003

The mechanisms controlling differentiation and lineage specification of neural stem cells are still poorlyunderstood, and many of the molecules involved in this process and their specific functions are yet unknown.We investigated the effect of apoptosis signal-regulating kinase 1 (ASK1) on neural stem cells by infecting adulthippocampus-derived rat progenitors with an adenovirus encoding the constitutively active form of ASK1.Following ASK1 overexpression, a significantly larger number of cells differentiated into neurons and asubstantial increase in Mash1 transcription was observed. Moreover, a marked depletion of glial cells wasobserved, persisting even after additional treatment of ASK1-infected cultures with potent glia inducers suchas leukemia inhibitory factor and bone morphogenetic protein. Analysis of the promoter for glial fibrillaryacidic protein revealed that ASK1 acts as a potent inhibitor of glial-specific gene transcription. However, thesignal transducers and activators of transcription 3 (STAT3)-binding site in the promoter was dispensable,while the activation of p38 mitogen-activated protein kinase was crucial for this effect, suggesting the presenceof a novel mechanism for the inhibition of glial differentiation.

The hippocampus is one of the two known areas of themammalian brain where neurogenesis persists in adulthood (3,11, 30). Multipotent progenitor cells are located in the sub-granular zone of the hippocampus, where they proliferate,survive, and migrate to the granular zone that is finally formedby terminally differentiated neurons. Differentiation of thesestem cells is orchestrated by various extrinsic and heritableintrinsic factors that determine proper spatial and temporalcellular development as well as their integration into the pre-existing architecture of the dentate gyrus.

In order to study the events and stimuli that control differ-entiation and lineage commitment, multipotent neural stemcells can be isolated and expanded in vitro from primary cul-tures after serial passages in the presence of either basic fibro-blast growth factor (bFGF) or epidermal growth factor (13, 17,19, 28). Adult hippocampus-derived rat progenitors (AHPs)are bFGF-dependent, self-renewing neural progenitor cellsthat have been shown to give rise to all cell types found in thecentral nervous system, such as neurons, astrocytes, and oligo-dendrocytes (27). Furthermore, grafting of AHPs into neuro-genic areas of the brain like the hippocampus, the rostralmigratory pathway, and the retina showed the ability of thesecells to adopt even nonhippocampus neuronal phenotypes de-pendent on the site of grafting, indicating that AHPs can notonly integrate well into the host tissue but also differentiateaccording to surrounding environmental cues (13, 35, 38, 42).

A variety of stimuli promoting neurogenesis from neuralstem cells in vivo and in vitro have been identified, such asretinoic acid, neurotrophins, steroid hormones, and insulin-like growth factor 1 (1, 5, 12, 37). Even though the exactmechanisms of action are still poorly understood, these neu-rogenic stimuli promote the expression of several basic helix-loop-helix transcription factors (bHLH) mediating neuronaldifferentiation, such as Mash1, neurogenin1 and 2, andNeuroD. Overexpression and loss-of-function studies havesuggested that precise temporal and spatial expression of thesebHLH transcription factors is critical for proper developmentof the nervous system (7, 8, 15, 34). While some members ofthe neural subclass of bHLH factors, such as NeuroD in post-mitotic neuronal precursors, are expressed at later stages ofneuronal differentiation, Mash1 and neurogenin1 are ex-pressed in proliferating neuronal progenitors and early differ-entiating neurons (9, 20, 21). Astroglial differentiation, how-ever, has been shown to be promoted by notch signaling (40) aswell as by members of the gp130 group of cytokines, including,among others, leukemia inhibitory factor (LIF) and ciliaryneurotrophic factor (2, 4, 17, 29). LIF activates a receptor-associated tyrosine kinase, the Janus kinase, which in turnphosphorylates STAT1 and STAT3. These cytoplasmic pro-teins dimerize upon phosphorylation and are translocated tothe nucleus after binding the CBP/p300 coactivator complex.Once inside the nucleus, they activate transcription of glialfibrillary acidic protein (GFAP), a glial-specific intermediatefilament, via the STAT1- and STAT3-binding site of the GFAPpromoter. Bone morphogenetic proteins (BMPs) have beenshown to be able to promote neuronal as well as astroglialdifferentiation in various culture systems (14, 22, 24, 43). It wasrecently suggested that the ability of BMPs to promote one

* Corresponding author. Mailing address: Department of MedicalCell Biology, Institute of Anatomy and Cell Biology, Goteborg Uni-versity, Box 420, SE-40530, Gothenburg, Sweden. Phone: 46 31 7733151. Fax: 46 31 773 3360. E-mail: [email protected].

† Present address: Department of Life Sciences, Graduate School ofArts and Sciences, University of Tokyo, Tokyo 153-8902, Japan.

280

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

rather than the other fate is dependent on the presence orabsence of neurogenin, since neurogenin has been shown tobind and thereby sequester the CBP/p300 coactivator complex(36). Neurogenin thereby actively inhibits the glia-inducingcapacities of LIF and BMPs by sequestering their coactivatorsaway from the glia-specific promoter.

Apoptosis signal-regulating kinase 1 (ASK1) is a well con-served and ubiquitously expressed intracellular serine/threo-nine mitogen-activated protein (MAP) kinase kinase kinasethat was first identified as a mediator of tumor necrosis factoralpha-induced apoptosis by activation of the p38 MAP kinaseand the c-jun N-terminal-activating kinase pathways (16).ASK1 is involved in induction of apoptosis in response tovarious other cytotoxic stresses, such as UV light, Fas, or re-active oxygen species in several cell types of neuronal origin,such as sympathetic neurons and differentiated PC12 cells (18).However, recent studies of keratinocytes and undifferentiatedPC12 cells showed that ASK1 could promote cellular differen-tiation and survival rather than apoptosis, suggesting thatASK1 might have opposing functions depending on cell typeand status (33, 39).

However, the cell types used in these two studies are eithernonneural cells or tumor cells of the peripheral nervous systemand therefore do not provide information about the role ofASK1 in the regulation of lineage determination in endoge-nous multipotent neural stem cells of the central nervous sys-tem (CNS).

In the present study, we used AHPs in order to investigatethe effect of ASK1 signaling on survival, proliferation, andlineage differentiation in progenitor cells of the CNS. We ad-dressed these questions because further understanding of therole of ASK1 signaling may be of potential value for futurecell-based therapeutic strategies in neurodegenerative dis-eases.

MATERIALS AND METHODS

Cell culture of AHPs. AHPs were kindly provided by F. Gage (Salk Institute,La Jolla, Calif.) and maintained as described previously (26). Clonal progenitorcells were used between passages 10 and 20 postcloning. For proliferating con-ditions, clonal progenitor cells were cultured in Dulbecco’s modified Eagle’smedium–Ham’s F12 (1:1) containing N2 supplement (Life Technologies), L-glutamine, and 20 ng of recombinant human bFGF per ml. When used forexperiments, the cells were plated at different densities on polyornithine-laminin-coated plates in medium lacking the mitogen bFGF.

Adenoviral vectors. The recombinant adenoviruses encoding LacZ (Ad-LacZ),the hemagglutinin (HA)-tagged kinase mutant ASK1 (Ad-ASK1-KM), and theconstitutively active ASK1 (Ad-ASK1-�N) were described previously (23, 31,32). The adenovirus was propagated in 293 cells to produce high-titer stocks. Thecells were infected for 3 to 4 days, harvested, and lysed by five cycles of freeze-thawing. The lysate was centrifuged at 10,000 � g for 20 min at 4°C, and thestocks were kept at �80°C. Infection efficiency was judged by staining for �-ga-lactosidase.

Growth factor treatment. For treatment with LIF or BMP, AHPs were cul-tured as described and infected with Ad-ASK1-�N, Ad-ASK1-KM, or Ad-LacZon days in vitro 2 (DIV 2). After another 24 h, the cells were treated with eitherLIF (GIBCO BRL) or BMP6 (kindly provided by Creative Biomolecules) untilfurther analysis. Concentrations ranged from 30 to 50 ng/ml for BMP6 and from50 to 80 ng/ml for LIF in order to induce expression of GFAP, since theseconcentrations have been shown to effectively induce GFAP expression in vari-ous culture systems (6, 24, 25, 36).

Treatment with SB 203580 and SB 202190. SB 203580 and SB 202190 werepurchased from Calbiochem, dissolved in dimethyl sulfoxide (DMSO) at 20 mM,and stored at �20°C. At the time of treatment, stocks of the inhibitor werethawed, prediluted in medium, and added to the well to the designated final

concentration. For long-term analysis, the inhibitor was repeatedly added to theculture medium every 2 days.

Immunohistochemistry. (i) 3,3-Diaminobenzidine staining. AHPs were cul-tured on chamber slides (Labtek; Nunc) and were fixed in 4% paraformaldehydefor 20 min at room temperature. After being washed with phosphate-bufferedsaline (PBS), the cells were quenched with 3% hydrogen peroxide for 10 min,washed again with PBS, and preincubated with Tris-buffered saline (TBS)-basedblocking buffer containing 0.1% Triton X-100, 3% bovine serum albumin, and3% normal horse serum for 1 h at room temperature. The cultures were thenincubated with mouse monoclonal antimicrotubule-associated protein 2ab(Map2ab) (Sigma-Aldrich) diluted 1:1,000 in the same blocking buffer. Afterthree washes in TBS, the cultures were incubated for 1 h at room temperaturewith biotinylated anti-mouse antibody (Vector Laboratories) diluted 1:200. Thecells were washed, and labeling was visualized by incubation with an ABC Elitekit (Vector Laboratories) for 1 h at room temperature and finally with 0.05%3,3-diaminobenzidine–0.03% H2O2 for 5 min.

(ii) Immunofluorescence. For immunofluorescence, the cells were cultured onchamber slides and fixed for 15 min with ice-cold methanol 4 and 9 days afterinfection. The cells were incubated in TBS-based blocking buffer for 1 h at roomtemperature. After being rinsed with PBS and TBS, the cells were incubated withprimary antibodies diluted in the same blocking buffer overnight at 4°C. Afterthree washes with PBS and TBS, the cells were incubated with fluoresceinisothiocyanate- or Texas red-conjugated secondary antibodies (Jackson Immu-noresearch) diluted 1:100. For nuclear counterstaining, the cells were incubatedin 50 ng of Hoechst 33258 (Sigma-Aldrich) per ml for 20 min before beingmounted in Dako fluorescent medium (Dakoplatts AB). The following primaryantibodies were used: mouse monoclonal anti-Map2ab (1:100), mouse monoclo-nal anti-GalC (1:100; Boehringer Mannheim), and rabbit polyclonal anti-GFAP(1:400; Dako).

Neurite outgrowth assay. For neurite outgrowth assay, AHPs were cultured onchamber slides and infected with adenoviral constructs carrying the ASK1 kinasemutant, ASK1-KM, or the constitutively active ASK1, ASK1-�N, as described.After another 5 days, the cells were fixed and stained for Map2ab and observedunder a standard light microscope. To determine the percentage of neurite-bearing cells, the number of Map2ab-positive cells with a process longer than onecell diameter was determined and compared to the total number of Map2ab-positive cells.

[3H]thymidine incorporation assay. For [3H]thymidine incorporation assays,AHPs were cultured on 24-well plates as described and infected with Ad-ASK1-�N, Ad-ASK1-KM, or Ad-LacZ. At the time of analysis, the cells were pulsedwith 0.5 �Ci of [methyl-3H]thymidine (Amersham) per ml of medium for another4 h. Then the cells were washed once with ice-cold PBS and precipitated with 5%trichloracetate on ice for 30 min. The precipitate was then dissolved in 1 MNaOH, neutralized with 1 M HCl, and transferred to vials containing 10 ml ofscintillation fluid. Radioactivity was counted by using a �-counter (Wallac).

MTT assay. The MTT assay is based on the ability of mitochondria of viablecells to reduce 3-(4,5-dimethyldiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT), a light yellow compound, into a dark blue product, formazan MTT, thatcan be quantified spectrophotometrically. In order to assess cell survival, cellswere cultured on 24-well plates as described. Sixty microliters of MTT solution(5 mg/ml; Sigma) was added to the medium, and the mixture was cultured foranother 4 h at 37°C. Solubilization buffer was then added, and the mixture wascultured for another 30 min to release the formazan salt. Then absorbance wasmeasured at 570 nm.

LDH cytotoxicity assay. Cell death was determined by a method based onspectrophotometric measurement of lactate dehydrogenase (LDH) release byusing the Cyto Tox 96 nonradioactive cyotoxicity assay kit (Promega) accordingto the manufacturer’s protocol. For each treatment group, total LDH release wasdetermined after addition of Triton X-100 to a final concentration of 10%. Theresults in Ad-ASK1-�N-infected wells were expressed as percentages relative tothe maximal LDH release in each treatment group and compared to the ratio forthe Ad-ASK1-KM-infected control.

Fluorescence activated cell sorter (FACS) analysis. Propidium iodide (PI)staining was performed using the Cycle TEST PLUS DNA reagent kit (BectonDickinson) according to the manufacturer’s protocol. Briefly, cells were seededon 6-well plates and infected with Ad-ASK1-�N or Ad-ASK1-KM on DIV 2 asdescribed above. After 4 days of infection, the cell nuclei were isolated andstained with PI before the emitted fluorescence (25,000 nuclei/sample) wasanalyzed using a FACS Calibur flow cytometer (Becton Dickinson).

Luciferase reporter assay. For luciferase reporter assays, 35,000 cells wereseeded on 12-well plates as described. The human GFAP promoter (positions�1,873 to �130) and the STAT3-binding site mutated human GFAP promoterwere gifts from T. Honjo (Kyoto University, Japan) and were described previ-

VOL. 24, 2004 EFFECT OF ASK1 ON NEURONAL AND ASTROGLIAL DEVELOPMENT 281

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

ously (40). The Mash1 promoter was provided by J. E. Johnson (University ofTexas Southwestern Medical Center) (41), and the p38 MAP kinase expressionvectors for the wild-type and dominant-negative mutant were donated by M.Landstrom (Ludwig Institute for Cancer Research, Uppsala, Sweden). Afteranother 2 days, the cells were transfected with the reporter plasmids and expres-sion vectors by using FuGENE6 (Boehringer Mannheim). The total amount ofDNA was kept constant. The cells were infected 4 h after transfection withadenovirus encoding LacZ, ASK1-�N, or ASK1-KM as described. In the case oftreatment with LIF or BMP, AHPs were treated with the growth factor 24 h afterinfection and a luciferase assay was performed after another 12 h. Absoluteluciferase activity was standardized to the total protein amount or �-galactosi-dase-activity. In each experiment, the samples were analyzed in triplicate, andeach experiment was repeated at least three times.

Western blotting. For Western blotting analysis, AHPs were cultured on poly-ornithine-laminin-coated 24-well plates as described. At the time of harvest, thecells were washed with PBS and lysed on ice in buffer containing 150 mM NaCl,50 mM Tris-HCl (pH 8.0), 0.5 mM EDTA, 1% NP-40, 0.5% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-l-propanesulfonate}, and 0.1% sodiumdodecyl sulfate (SDS). Phenylmethylsulphonyl fluoride and aprotinin werefreshly added each time to a final concentration of 1 mM to inhibit proteaseactivity. Cellular debris was removed by centrifugation, and the protein concen-tration of each sample was determined by using a DC protein assay (Bio-Rad).Equal amounts of protein were separated on SDS–10% to 12% polyacrylamidegel electrophoresis and then electroblotted onto a polyvinylidene difluoridemembrane by semidry transfer (Bio-Rad). After being incubated with 5% nonfatskim milk in TBS with 0.1% Tween 20, the membranes were probed with primaryantibodies. The following primary antibodies were used: rabbit polyclonal anti-GFAP (1:2,000; Dako), mouse monoclonal anti-�-III-tubulin (1:500; Sigma),mouse monoclonal antiactin (1:500; Sigma), rabbit polyclonal anti-STAT3 andanti-phospho-tyr-STAT3 (1:1,000; Cell Signaling Technology, Inc.), rabbit poly-clonal antibodies against phosphorylated and total p38 MAP kinase (1:1,000;New England Biolabs, Inc.), and rabbit polyclonal anti-phospho-Smad1/5 (1:2,000; a gift from C. H. Heldin at Ludwig Institute, Uppsala, Sweden). The ratmonoclonal antibody against the HA tag (3F10; Roche) was directly conjugatedwith horseradish peroxidase and used at a concentration of 1:500. Horseradishperoxidase-conjugated anti-rabbit and anti-mouse immunoglobulin G were usedas secondary antibodies at concentrations of 1:5,000 (New England Biolabs).Signals were detected by enhanced chemiluminescence (ECL; New EnglandBiolabs).

RT PCR. Total RNA was extracted by the guanidinium thiocyanate method ofChomczynski and Sacchi (10). For cDNA synthesis, random hexamer primers(Roche) were used. First-strand cDNAs were synthesized from 1 �g of RNAusing Moloney murine leukemia virus reverse transcriptase (RT) (Invitrogen).The PCR was carried out by using standard protocol with Taq polymerase(Fermentas). The cycling parameters were as follows: denaturation at 94°C for 4min, followed by 25 to 28 cycles of denaturation at 94°C for 30 s, an annealingstep for 45 s, and 72°C for 1 min 30 s. The last cycle consisted of a final extensionstep at 72°C for 6 min. The following nucleotide primers were used: Mash1 sense,5�-GGA ACT GAT GCG CTG CAA AC, and antisense, 5�-CCT GCT TCCAAA GTC CAT TCC; and Actin sense, 5�-AAG ATG ACC CAG ATC ATGTTT GAG, and antisense, 5�-AGG AGG AGC AAT GAT CTT GAT CTT. Thesamples were run on 1% agarose gel containing ethidium bromide and analyzedby using a FLA 2000 plate reader (Fujifilm).

Microscopic analysis. The cultures were analyzed by standard light and epi-illumination fluorescence microscopy, and color images were generated by usinga Zeiss fluorescence microscope attached to a Nikon digital camera. The imageswere then processed for publication by using Adobe Photoshop (Adobe Sys-tems).

Statistics. All experiments were replicated at least three or four times. Overallsignificance was determined by submitting the data to one-way analysis of vari-ance (ANOVA). The significance of between-group differences was determinedby the Scheffe post hoc test.

RESULTS

Adenovirus-mediated ASK1 transduction yields efficient ex-pression in AHPs. Two days after being seeded, AHPs wereinfected with Ad-LacZ, Ad-ASK-KM, or Ad-ASK1-�N, whichis constitutively active because it lacks the N-terminally locatedbinding site for thioredoxin, the physiological inhibitor ofASK1 (32). Our adenoviral expression system provided infec-

tion efficiency up to 90% to 95%, as determined by staining for�-galactosidase (data not shown). Western blotting for the HAtag showed that expression of the infected constructs was al-ready detectable 8 h after infection. However, the expressionmaximum was not reached before 1 to 2 days after infectionand persisted until about 4 days after infection. Analysis of HAexpression for time points later than 4 days after infectionrevealed a gradual decrease of overexpressed ASK1-�N pro-tein (Fig. 1).

Moderate expression of ASK1 induces differentiation ofAHPs. In order to investigate the effect of ASK1 on the growthrate and cell survival of AHPs, cells were cultured as describedand infected with Ad-LacZ, Ad-ASK1-KM, or Ad-ASK1-�Non DIV 2. MTT assays, [3H]thymidine incorporation assays,LDH cytotoxicity assays, and FACS analysis were performedon day 2 or 4 after infection. While the MTT assay, an assayreflecting cell viability, already showed a strong decrease in cellviability 48 h after infection at a multiplicity of infection (MOI)of 100 and above, no significant reduction in cell viability wasobserved at MOIs of between 5 and 25 (Fig. 2A).

[3H]thymidine incorporation assays, however, which reflectcell proliferation, displayed a moderate but significant de-crease after 48 h of infection with Ad-ASK1-�N, even betweenMOIs of 10 and 25, compared to cell proliferation levels forAd-LacZ- or Ad-ASK1-KM-infected control cultures (Fig.2B). Considering that cell viability at these MOIs was unaf-fected, this result suggests that upon infection with Ad-ASK1-�N, AHPs were induced to differentiate.

The absence of significant cell death after infection withAd-ASK1-�N at MOIs between 5 and 25 was confirmed by theLDH cytotoxicity assay. However, cultures infected with Ad-ASK1-�N at MOIs of 100 and above displayed a significantincrease in LDH release compared to that of the Ad-ASK1-KM-infected control, confirming the notion that cell death wasinduced at these MOIs, as suggested by the MTT data (Fig.2C).

The nuclei of Ad-ASK1-infected AHPs were stained for PIand analyzed by FACS at 4 days after infection with Ad-LacZ,Ad-ASK1-KM, or Ad-ASK1-�N. A significantly higher num-ber of cells was arrested in G2/M phase upon infection withAd-ASK1-�N than for cultures infected with Ad-ASK1-KM(Fig. 2D). This result was consistently observed at MOIs be-tween 5 and 25. Ad-LacZ-infected cultures did not differ intheir staining profiles from those of Ad-ASK1-KM-infectedcultures at any tested MOI (data not shown). These findingsstrongly suggest that growth arrest was associated with differ-entiation of AHPs in response to moderate levels of ASK1overexpression.

FIG. 1. Adenovirus-mediated expression of ASK1-�N in AHPs.AHPs were infected with adenovirus carrying ASK1-�N on DIV 2 andharvested for Western blot analysis at the indicated time points.

282 FAIGLE ET AL. MOL. CELL. BIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

FIG. 2. ASK1 differentiates AHPs. (A) AHPs were infected with adenovirus carrying LacZ, ASK1-KM, or ASK1-�N on DIV 2, and afteranother 2 days, cells were subjected to MTT assay. (B) [3H]thymidine incorporation assays were done on cells as described for panel A. (C) AnLDH cytotoxicity assay was performed after 4 days of infection. (D) For FACS analysis, AHPs were cultured on 6-well plates, and PI staining wasperformed after 2 days of infection. The values represent means plus standard errors of the means (SEM; n � 3) from three independentexperiments. An asterisk indicates a P value of 0.05, two asterisks indicate a P value of 0.01, and three asterisks indicate a P value of 0.001in comparison to the LacZ-infected (A and B) or KM-infected (C) control. Results were determined by ANOVA and the Scheffe post hoc test.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

ASK1 promotes neuronal differentiation in adult hippocam-pal progenitors. In order to evaluate the effect of ASK1 ondifferentiation in AHPs, cells were infected with differentMOIs of Ad-ASK1-�N at 48 h after seeding and processed forimmunohistochemical staining after another 4 or 9 days. Cellswere stained for Map2ab, a marker for mature neurons, inorder to test the hypothesis that infection with Ad-ASK1-�Ninduces neuronal lineage determination in the progenitors. Incomparison to that for Ad-LacZ- or Ad-ASK1-KM-infectedcultures, the number of Map2ab-positive cells was found to beslightly increased 4 days after infection with Ad-ASK1-�N.However, a significant increase in Map2ab-positive cells wasnot observed before 8 to 9 days after Ad-ASK1-�N infection.This effect was already apparent at a MOI of 5 but was mostprominent at the MOIs of 10 and 25, where the number ofMap2ab-positive cells in Ad-ASK1-�N-infected cultures wasincreased about two- to threefold compared to that for thecontrols (Fig. 3A and B). Also, the number, length, and thick-ness of the processes of the MAP2ab-positive cells were in-creased in Ad-ASK1-�N-infected cultures compared to thoseof the control cultures. In order to quantify the neurite-induc-ing capacity of ASK1 in cultures of AHPs, we attempted toassess neurite outgrowth. AHPs were cultured on chamberslides as described, fixed, and stained for Map2ab after 5 daysof infection. When neurons exhibiting one or more neuriteslonger than one cell in diameter were counted and related tothe total number of Map2ab-positive cells, the percentage ofneurite-bearing cells was increased up to 52% � 7% in Ad-ASK1-�N-infected cultures, whereas only 21% � 5% of theneurons displayed neurite outgrowth in cultures infected withAd-ASK1-KM, suggesting that the kinase activity of ASK1plays an important role in the induction of neurite outgrowthin AHP cultures (Fig. 3C and D).

To further confirm the induction of neuronal differentiationby ASK1, we performed Western blot analysis in order toexamine the expression of �-III-tubulin, a neuron-specific iso-form of �-tubulin. As expected, we found significantly higherexpression of �-III-tubulin in Ad-ASK1-�N-infected culturesthan in Ad-LacZ- or Ad-ASK1-KM-infected cultures after in-fection with MOIs of between 5 and 25, indicating that theprogenitor cells adopted the neuronal phenotype in responseto ASK1 (Fig. 3E). Although this effect had already becomeapparent 4 days after infection, �-III-tubulin expression wasmaximal at 8 to 10 days after infection, reflecting the timeneeded for the progenitors to differentiate to neurons in thisculture system. In order to quantify the increase in �-III-tubu-lin expression, we performed densitometric analysis on threeindependent blots where �-III-tubulin signals were normalizedto the levels of �-actin present in the same sample and ex-pressed as percentages of the control sample infected withAd-LacZ. Protein levels of �-III-tubulin were consistentlyhigher in cultures infected with Ad-ASK1-�N than in the con-trol cultures, confirming the previous findings that overexpres-sion of Ad-ASK1-�N increases the expression of neuronalmarkers in cultures of AHPs (Fig. 3F). Furthermore, we ob-served that transcription of Mash1, a proneural gene, wasstrongly induced in AHPs in response to Ad-ASK1-�N over-expression compared to that for the control cultures. WhenAHPs were transfected with a 1.0-kb Mash1 promoter, lucif-erase activity was about two to three times higher in Ad-ASK1-

�N-infected cultures than in cultures infected with Ad-ASK1-KM or Ad-LacZ (Fig. 3G). In order to evaluate whetherAd-ASK1-�N overexpression actually increases endogenousMash1 mRNA expression in cultures of AHPs, we performedRT PCR analysis. We found that the progenitors indeed dis-played higher levels of mRNA for Mash1 after infection withAd-ASK1-�N than did the control cultures (Fig. 3H).

ASK1 promotes neuronal differentiation in a p38-dependentmanner. Since ASK1 is an upstream kinase in many knownsignaling pathways, we investigated which downstream kinasemight be the mediator of the ASK1-induced neuronal differ-entiation observed in our cultures. p38 MAP kinase has beenshown to be one of the most prominent and best characterizeddownstream effector kinases of ASK1 in many systems. Inorder to evaluate whether overexpression of Ad-ASK1-�N wasable to induce activation of the p38 MAP kinase pathway incultures of AHPs, cells were infected with adenovirus encodingLacZ, ASK1-KM, or ASK1-�N on DIV 2 as described. West-ern blot analysis for phosphorylated p38 MAP kinase was per-formed after another 24 h. While phosphorylated p38 MAPkinase was clearly detectable upon overexpression of Ad-ASK1-�N, detectable levels of activated p38 MAP kinase wereabsent in Ad-LacZ- and Ad-ASK1-KM-infected cultures (Fig.4A). When cultures were preincubated with the p38-specificinhibitor SB 203580, infected as described, and processed byimmunohistochemistry for Map2ab after 9 days of infection,Ad-ASK1-�N-infected cultures displayed a significantly highernumber of cells staining positive for Map2ab than Ad-ASK1-KM-infected cultures, a finding that is consistent with ourobservations described in Fig. 3. However, when Ad-ASK1-�N-infected cultures had been preincubated with SB 203580prior to infection, the number of Map2ab-positive cells did notsignificantly differ from that for the control cultures, suggestingthat inhibition of the p38 Map kinase pathway by SB 203580abolished the Ad-ASK1-�N-mediated induction of neuronaldifferentiation in our cultures (Fig. 4B and C). Furthermore,we sought to determine whether the Ad-ASK1-�N-mediatedinduction of the Mash1 gene was dependent on p38 MAPkinase activation. Therefore, AHPs were cultured as describedand transfected with a 1.0-kb Mash1 promoter on DIV 2. Priorto infection, cultures were incubated with either DMSO, SB203580, or SB 202190, and luciferase activity was determinedafter another 36 h. While Mash1 luciferase activity did notsignificantly change upon the presence of SB 203580 or SB202190 in cultures infected with Ad-ASK1-KM, the substantialincrease in luciferase activity observed in DMSO-treated Ad-ASK1-�N-infected cultures was found to be almost completelyabolished in the presence of either one of the two p38 MAPkinase-specific inhibitors (Fig. 4D).

ASK1 inhibits the development of glial cells in cultures ofAHPs. In order to determine possible effects of ASK1 on theglial cell population in the cultures, cells were infected withadenovirus encoding LacZ, ASK1-KM, or ASK1-�N on DIV 2as described. After another 4 or 10 days in culture, cells wereharvested, lysed, and subjected to Western blot analysis forGFAP. When examined at 4 days after infection, GFAP ex-pression was already strongly decreased at a MOI of 5, and itdecreased further in an MOI-dependent manner than did theAd-LacZ- or Ad-ASK1-KM-infected cultures (Fig. 5A). In or-der to quantify this MOI-dependent decrease of GFAP expres-


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

FIG. 3. ASK1 induces neuronal differentiation in AHPs. (A and B) AHPs were infected with adenovirus encoding LacZ, ASK1-KM, orASK1-�N and processed for immunofluorescent staining for Map2ab 9 days after infection. (C and D) A neurite outgrowth assay was performedat 5 days after infection. (E) AHPs were immunoblotted for �-III-tubulin at 9 days after infection. (F) Densitometric analysis of three independentWestern blots for �-III-tubulin is shown. (G) AHPs were transfected with Mash1-Luc 4 h prior to infection. Luciferase activity was determinedat 36 h after transfection. (H) RT PCR analysis of AHPs was performed 2 days after infection with Ad-LacZ, Ad-ASK1-KM, or Ad-ASK1-�N.The values represent means � SEM (n � 3) from three independent experiments. An asterisk indicates a P value of 0.05, two asterisks indicatea P value of 0.01, and three asterisks indicate a P value of 0.001 in comparison to the LacZ-infected (B, F, and G) or KM-infected (D) control.Results were determined by ANOVA and the Scheffe post hoc test. Scale bar, 10 �M.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

sion, we performed densitometric analysis on three indepen-dent blots as described above. Already having an MOI of 5, theexpression of GFAP decreased to 25% at 4 days after infectionwith Ad-ASK1-�N, decreasing further to 10% to an MOI of50, which appeared to be the most effective MOI in this respect(Fig. 5B).

Immunohistochemical staining for GFAP revealed that, in-deed, the number of GFAP-positive cells was significantly de-creased in Ad-ASK1-�N-infected cultures of AHPs comparedto that for Ad-LacZ- or Ad-ASK1-KM-infected cultures after4 days of infection (Fig. 5C). Counting of GFAP-positive cellsat this time point confirmed that only 7% � 2% of all cellswere positive for GFAP in Ad-ASK1-�N-infected cultures,whereas in control cultures 21% � 4% of the cells were GFAPpositive (Fig. 5D). However, when cells were left in culture for9 to 10 days, GFAP expression in Ad-ASK1-�N-infected cul-tures was increased again, even reaching baseline levels. Thisresult was determined by immunohistochemical staining aswell as by Western blotting (Fig. 5D and E).

Immunohistochemical staining for GalC, a marker for oli-godendroglia, did not show any difference in the number ofoligodendrocytes in cultures infected with Ad-ASK1-�N thanin Ad-ASK1-KM-infected cultures, excluding the possibilitythat the progenitors adopted oligodendroglial instead of astro-glial fate (data not shown).

ASK1 decreases the astroglial cell population by repressingGFAP promoter activity in AHPs. In order to investigatewhether the glia-depleting effect of ASK1 on cultures of AHPsis due to direct inhibition of GFAP promoter activity, AHPswere transfected with a luciferase vector controlled by thehuman GFAP (hGFAP) promoter at 4 h prior to infection withAd-ASK1. Luciferase activity was measured 36 h after infec-tion. Overexpression of ASK1-�N efficiently suppressedhGFAP promoter activity in an MOI-dependent manner com-pared to that for Ad-LacZ-infected cultures, suggesting thatASK1 directly targets the GFAP promoter (Fig. 6). Further-more, it can be observed that ASK1-KM significantly increasesGFAP promoter activity, suggesting the presence of endog-enously activated ASK1 that might control GFAP promoteractivity in native cultures of AHPs. However, no significantincrease in protein levels of GFAP could be observed in cul-tures infected with Ad-ASK1-KM.

ASK1 acts as an efficient inhibitor of glial cell developmenteven in the presence of the glia-instructive signals of LIF andBMP. LIF and BMP are known to be strong inducers of GFAPin progenitor cells of various origins by acting instructively on

FIG. 4. ASK1-induced neuronal differentiation is dependent onp38 MAP kinase activation. (A) AHPs were infected with adenoviruscarrying LacZ, ASK1-KM, or ASK1-�N on DIV 2 and harvested forWestern blot analysis for phosphorylated p38 after another 24 h.(B) AHPs were incubated with 10 �M SB 203580 prior to infectionwith Ad-ASK1-KM or Ad-ASK1-�N. The culture medium was sup-

plemented with SB 203580 every second day, and cells were processedfor immunohistochemical staining for MAP2ab 9 days after infection.(C) Cell counts of Map2ab-positive cells cultured as described forpanel B. (D) AHPs were transfected with Mash1-Luc and incubatedwith DMSO as a control or with SB 203580 or SB 202190 prior toinfection with Ad-ASK1-KM or Ad-ASK1-�N. Luciferase activity wasmeasured 36 h after infection. The values represent means � SEM (n� 3). Three asterisks indicate a P value of 0.001 in comparison to theKM-infected control, two number signs indicate a P value of 0.01,and three number signs indicate a P value of 0.001 in comparison tothe DMSO control infected with the same viral construct. Results weredetermined by ANOVA and the Scheffe post hoc test. Scale bar, 10�M.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

the determination of astroglial cell fate. In order to determinewhether the decrease of GFAP expression in AHPs culturesafter infection with Ad-ASK1-�N can be restored by glia-inducing signals, LIF or BMP6 was added to the cultures at24 h after infection at a concentration of 50 ng/ml and 30ng/ml, respectively. Western blot analysis for GFAP was per-formed after 3 more days. As expected, LIF and BMP6 in-duced a strong increase in GFAP expression in control culturesinfected with Ad-ASK1-KM. However, when added to Ad-ASK1-�N-overexpressing cultures, LIF and BMP6 were not

potent enough to increase GFAP to baseline levels, unlike theuntreated culture infected with Ad-ASK1-KM, even thoughGFAP expression was slightly stronger than in untreated cul-tures infected with Ad-ASK1-�N (Fig. 7A). We did not ob-serve any differences in GFAP expression between Ad-LacZ-and Ad-ASK1-KM-infected cultures treated with the cyto-kines. This finding suggests that even strong glia inducers, suchas LIF and BMP, are not able to restore GFAP levels incultures that are depleted of GFAP-positive cells by ASK1 tothe level observed in the control cultures. Densitometric anal-

FIG. 5. ASK1 inhibits glial differentiation in AHPs. (A) AHPs were cultured without bFGF for 2 days and then infected with adenoviruscarrying LacZ, ASK1-KM, or ASK1-�N. After another 4 days, cells were processed for Western blot analysis. (B) Densitometric analysis of threeindependent Western blot assays for GFAP is shown. (C) Immunofluorescence staining for GFAP 4 days after infection. (D) Cell counts ofGFAP-positive cells stained 4 and 9 days after infection. (E) AHPs were cultured for 10 days after infection and immunoblotted for GFAP. Thevalues represent means � SEM (n � 3). Two asterisks indicate a P value of 0.01, and three asterisks indicate a P value of 0.001 in comparisonto the LacZ-infected control. Results were determined by ANOVA and the post hoc Scheffe test. Scale bar, 10 �M.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

ysis of three Western blots from separate experiments con-firmed this observation. Hence, LIF and BMP6 are able toincrease GFAP expression in the presence of Ad-ASK1-�Nbut to only a limited extent (Fig. 7B).

In order to determine the numbers of GFAP-positive cellsupon exposure to LIF and BMP, AHPs were infected on DIV2 as described and treated with one of the cytokines at 24 hafter infection. After another 3 days, cells were fixed and pro-cessed for immunohistochemical staining for GFAP. While22% 3% of the cells stained positive for GFAP in untreatedAd-ASK1-KM-infected cultures, addition of 50 ng of LIF or 30ng of BMP per ml resulted in an increase of GFAP-positivecells up to 52% 1% and 38% 2%. When Ad-ASK1-�N-infected cultures were left untreated, the number of GFAP-positive cells was reduced to 9% 2%, a phenomenon exten-sively described in Fig. 5. However, when LIF or BMP wasadded to Ad-ASK1-�N-overexpressing cultures, these cyto-kines were hardly able to elevate the numbers of GFAP-posi-tive cells to baseline levels, unlike the results for untreatedcultures infected with Ad-ASK1-KM. This finding gives furthersupport to the hypothesis suggesting that ASK1 is an importantand potent inhibitor of glial cell development (Fig. 7C).

Furthermore, we examined whether LIF and BMP6 hadsimilar effects on the GFAP promoter activity in Ad-ASK1-�N-infected progenitor cultures. Cells were transfected withhGFAP promoter luciferase plasmid 4 h prior to infection withAd-ASK1-�N and treated with 80 ng of LIF or 50 ng of BMP6per ml after another 24 h. Luciferase activity was then mea-sured at 12 h after addition of the growth factor. As alreadyshown in Fig. 4, ASK1-�N induced an MOI-dependent de-crease in GFAP promoter activity of up to 20% to 25% at anMOI of 25 compared to that for untreated Ad-LacZ-infectedcultures. Treatment with LIF or BMP increased GFAP pro-moter activity significantly in Ad-LacZ- and Ad-ASK1-KM-infected wells as well as in cultures infected with Ad-ASK1-�Nat an MOI of 5. However, at an MOI of 25, LIF and BMP6failed to significantly increase GFAP levels in Ad-ASK1-�N-

infected cultures. Compared to the LIF- and BMP-treatedcontrols, Ad-ASK1-�N-infected cultures that had received LIFor BMP treatment consistently showed a significant decreasein GFAP promoter activity (Fig. 7D). Interestingly, when com-paring the potency of the applied growth factors, we found thatLIF appeared to be more potent than BMP6 in increasingprotein levels of GFAP, whereas on the level of GFAP pro-moter activity, BMP6 seemed to be more effective. Takentogether, these data show that the strong reduction of GFAPprotein levels in Ad-ASK1-�N-infected cultures of AHPs isdue to direct inhibition of GFAP promoter activity. This re-duction could not be restored by instructive glia-inducing sig-nals, such as LIF and BMP, either on the level of promoteractivity or on the protein level.

ASK1 decreases GFAP protein in cultures of AHPs inde-pendently of the STAT3 pathway. In order to investigatewhether the inhibition of GFAP promoter activity by ASK1 isdependent on the STAT3-binding site of the promoter orwhether ASK1 exerts its action by interfering with known up-stream pathways of the STAT3-binding site, we transfectedAHPs with a STAT3-binding site mutant hGFAP luciferaseconstruct (STAT3mut-Luc), disrupting the binding of STAT3to the GFAP promoter, and measured luciferase activity 36 hafter infection with Ad-ASK1. Ad-LacZ-infected culturestransfected with STAT3mut-Luc showed a significant decreasein luciferase activity compared to that for cultures transfectedwith the native hGFAP promoter, suggesting that glial celldevelopment in AHPs is dependent on STAT3 activation byendogenous signals. However, ASK1-�N was still able to re-press the activity of the STAT3mut-Luc with the same effi-ciency as in cultures transfected with the native hGFAP-Luc,suggesting that negative regulation of the GFAP promoter byASK1 is independent of the STAT3-binding site (Fig. 8A).Furthermore, we investigated whether the levels of phosphor-ylated Smad and tyrosine-phosphorylated STAT3 werechanged in response to ASK1-�N overexpression. As ex-pected, no changes in phosphorylation levels of either STAT3or Smad were found, further supporting the finding that ASK1-mediated inhibition of GFAP promoter activity is indeed in-dependent of activation of the STAT3-binding site in theGFAP promoter (Fig. 8B and C).

Repression of GFAP promoter activity by ASK1 is mediatedvia the p38 MAP kinase pathway. In order to determinewhether p38 MAP kinase signaling is involved in mediatingGFAP promoter repression in AHPs by ASK1, cells weretransfected with hGFAP-Luc and infected with Ad-LacZ, Ad-ASK1-KM, or Ad-ASK1-�N as described before. They werethen treated with 10 �M SB 203580 at the time of infection,and luciferase activity was determined after another 36 h. Fig-ure 9A shows that treatment with the p38 MAP kinase inhib-itor did not change luciferase activity in Ad-LacZ- or Ad-ASK1-KM-infected cultures. In contrast, GFAP promoteractivity was significantly increased in Ad-ASK1-�N-infectedcultures after treatment with SB 203580 compared to that forAd-ASK1-�N-infected cultures that received DMSO as a con-trol. This finding suggests that the ASK1-induced repression ofthe GFAP promoter is indeed mediated via the p38 MAPkinase. In order to confirm this finding, we overexpressedASK1-�N and p38 dominant negative (p38DN) and cotrans-fected with hGFAP-Luc. Luciferase activity was analyzed after

FIG. 6. ASK1 signaling inhibits GFAP promoter activity in AHPs.AHPs were infected on DIV 2 with adenovirus carrying LacZ, ASK1-KM, or ASK1-�N and transfected with hGFAP-Luc 4 h prior toinfection. Luciferase activity was measured after another 36 h. Thevalues represent the means � SEM (n � 4) from one of three inde-pendent experiments. Three asterisks indicate P values of 0.001 incomparison to the LacZ-infected control. Results were determined byANOVA and the Scheffe post hoc test.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

another 36 h as described before. We observed that cotrans-fection with p38DN significantly inhibited the repressive effectof ASK1 on the GFAP promoter compared to that for ASK1-�N-transfected cultures that were cotransfected with eitherempty vector or the p38 wild type (Fig. 9B). In order to deter-mine whether inhibition of the p38 MAP kinase pathway priorto overexpression of Ad-ASK1-�N would restore protein lev-els of GFAP, Western blot analysis for GFAP was performedon Ad-LacZ- and Ad-ASK1-�N-infected cultures that hadbeen preincubated, respectively, with either DMSO as a con-trol or with 10 �M SB 203580. Figure 9C shows that GFAPprotein levels can indeed be restored upon overexpression ofAd-ASK1-�N when SB 203580 is present, while DMSO-treated control cultures infected with Ad-ASK1-�N show thealready described strong reduction in GFAP expression. Thesedata strongly suggest that ASK1-induced inhibition of GFAPexpression is mediated via the p38 MAP kinase pathway.

DISCUSSION

The role of ASK1 in cell fate specification. We have inves-tigated the role of ASK1 in survival, proliferation, and differ-entiation of CNS stem cells by using bFGF-dependent multi-potent neural progenitor cells derived from the adult rathippocampus. Our data show that after infection with Ad-ASK1-�N a significantly higher number of cells displayed neu-ronal features compared to that for control cultures. Eventhough our cultures contained a significant number of imma-ture neural progenitor cells, the presence of already prediffer-entiated precursors of the neuronal or the glial lineage at thetime of infection cannot be excluded. It therefore remains tobe investigated whether the increased number of neurons afterinfection with Ad-ASK1-�N is due to an induction of neuronalfate in neural stem cells or to promotion of neuronal differen-tiation in already predifferentiated precursors. In order to de-termine which of the different populations of precursors re-spond to ASK1, it would be necessary to distinguish betweenthem and analyze their responses to ASK1 separately. How-ever, determining the single differentiation stages between animmature neural stem cell and a terminally differentiated neu-ron is rather difficult, mainly due to the limited availability ofappropriate markers to distinguish these different develop-mental stages. Nonetheless, we observed a significant increasein Mash1 promoter activity in response to ASK1-�N in theprogenitor cultures. This finding indicates that there is a rea-son to believe that a substantial number of yet uncommittedprogenitors in the culture respond to ASK1 signaling by dif-ferentiating to the neuronal lineage, since Mash1 is one of theearliest markers expressed in committed neural progenitors.

There have been no previous investigations about the role ofASK1 in glial-cell development. However, our results indicate

FIG. 7. Effect of ASK1 on GFAP expression in the presence ofBMP and LIF. (A) AHPs were infected with Ad-ASK1-KM or Ad-ASK1-�N on DIV 2 and treated with 50 ng of LIF or 30 ng of BMP6per ml, respectively, after another 24 h. GFAP expression was deter-mined by Western blot analysis after another 3 days. (B) Densitometricanalysis of three independent Western blots. (C) Cell counts of GFAP-positive cells infected and treated as described for panel A. (D) AHPswere transfected with hGFAP-Luc 4 h prior to infection and exposedto 80 ng of LIF or 50 ng of BMP6 per ml after another 24 h. Luciferaseactivity was determined 12 h after addition of the growth factor. The

values represent means � SEM (n � 3). One asterisk indicates a Pvalue of 0.05, two asterisks indicate a P value of 0.01, and threeasterisks indicate a P value of 0.001 in comparison to the buffer-treated control. One number sign indicates a P value of 0.05, twonumber signs indicate a P value of 0.01, and three number signsindicate a P value of 0.001 in comparison to KM-infected (B and C)or LacZ-infected (D) controls treated with the same growth factor.Results were determined by ANOVA and the Scheffe post hoc test.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

that ASK1 may play an important role in negative regulation ofglial-cell development in cultures of adult neural progenitorcells. Extrinsic and intrinsic factors that influence the fate ofuncommitted multipotent progenitor cells can exert their ac-tion in two distinct ways. They can either instruct the progen-itors to commit to a particular lineage, or they can allow thecells to be more susceptible to other factors determining pro-liferation, survival, and lineage restriction. In order to definethe strength of ASK1-mediated inhibition of glial-cell devel-opment, we challenged the effect of ASK1 by simultaneousapplication of LIF or BMP, two known instructive inducers ofastroglial cell fate. We showed that the inhibition of glial-celldifferentiation by ASK1 is effective and robust, since neither ofthe two applied factors was able to restore either the expres-sion levels of GFAP protein or the number of GFAP-positivecells back to baseline levels. The amount of cell death was notsignificantly elevated in Ad-ASK1-�N-infected cultures at theMOIs used for analysis of GFAP expression, indicating thatthe reduction in GFAP protein was unlikely to be due to celldeath of GFAP-positive cells or astroglial precursors in ourcultures. Furthermore, the decrease of GFAP promoter activ-ity in response to ASK1 signaling supports the conclusion thatthe reduction in the number of GFAP-positive cells is indeeddue to an inhibition of glial-cell development rather than tocell death.

However, when infected cells were left in culture for longerperiods, the initially observed absence of GFAP-positive cellsafter infection with Ad-ASK1-�N was no longer apparent. Weconsidered two different scenarios to explain this phenomenon.Since the infection efficiency of our adenoviral transduction isbelow 100%, it might be possible that the population of initiallyuninfected cells expanded in comparison to the population ofinfected cells in the course of the culture. These cells couldthen develop into astroglial cells and thereby replace thosecells within the population of infected cells that were initiallymeant to adopt glial fate. Even though we cannot completelyexclude this possibility, we consider it unlikely that the popu-lation of initially uninfected cells can expand in such a tremen-dous way, especially since our infection efficiency was as highas 90% to 95% and would therefore leave a rather smallnumber of cells uninfected. It seems more likely that glialdifferentiation in neural stem cells or early glial precursors isinhibited by ASK1 as long as the adenoviral infection is able toprovide a sufficient expression of ASK1-�N protein in the cells.Since adenoviral DNA, in contrast to retroviral DNA, does notintegrate into the genome of the infected cell, the number ofvirus particles in each cell decreases with every cell division.Thus, the levels of ASK1-�N protein decrease in the course ofthe culture, and the inhibitory effect of ASK1 on glial differ-entiation in neural stem cells or early glial precursors willdecrease or even be abolished as soon as ASK1-�N proteinlevels drop beneath the threshold necessary to efficiently in-hibit glial differentiation. In such a case, the thereby disinhib-ited cells would be able to resume their development intoastroglial cells and the levels of GFAP protein would notdisplay any significant reduction compared to those of thecontrol cultures.

Mechanism of ASK1-mediated inhibition of glial-cell devel-opment. LIF and BMP are known to be instructive inducers ofastroglial cell fate in various systems. Since BMPs have also

FIG. 8. ASK1 inhibits GFAP promoter activity independently ofSTAT3 activation. (A) AHPs were transfected with hGFAP-Luc or aSTAT3-mutated hGFAP-Luc 4 h prior to infection with Ad-LacZ,Ad-ASK1-KM, or Ad-ASK1-�N. Luciferase activity was determinedafter another 36 h. (B) AHPs were treated with 80 ng of LIF or 50 ngof BMP6 per ml at 24 h after infection with ASK1-KM or ASK1-�N,respectively, and cells were harvested for Western blot analysis forphospho-tyr-STAT3 20 min after addition of the growth factor.(C) Analysis of Smad phosphorylation performed as described forpanel B. The values represent means � SEM (n � 4). An asteriskindicates a P value of 0.05, two asterisks indicate a P value of 0.01in comparison to the LacZ-infected control, and a number sign indi-cates a P value of 0.05 in comparison to the hGFAP-Luc-transfectedcontrol infected with the same construct. Results were determined byANOVA and the Scheffe post hoc test.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

been reported to promote neuronal differentiation, it was re-cently suggested that the amounts of endogenous neurogeninin each cell determine the response of the cell to BMP signal-ing (36). When levels of neurogenin are low, BMP inducesastroglial fate in immature neural progenitor cells. In our cul-ture system, LIF and BMP appeared to be very potent inducersof astroglial fate, suggesting that levels of endogenous neuro-genin are rather low. Nonetheless, it would be possible that theeffect of ASK1 on AHPs described in this study is mediated viainduction or activation of endogenous neurogenin, since theeffects of ASK1 on AHPs described in our study somewhatresemble the effects of neurogenin on stem cells reported else-where (36). However, it seems unlikely that neurogenin acts asa direct downstream mediator of ASK1 signaling in AHPs. It iswell known that instructive glia-inducing signals, such as LIF,CNTF, and BMP, activate the GFAP promoter via the bindingsite for STAT3, which is one of the major cytoplasmic signalingmolecules activated by LIF and CNTF. BMPs exert their ac-tion by phosphorylation of Smads, forming a complex withCBP/p300. This complex in turn binds to STAT3 and is therebysequestered to the STAT3-binding site of the GFAP promoter,activating the promoter.

In our study, however, we found that the STAT3-binding siteof the GFAP promoter was dispensable for the ASK1 effect,and we could not detect any changes in the phosphorylationstatus of STAT3 or Smad. This finding might present a yetunknown mechanism of inhibition of glial-cell developmentthat is independent from the mechanism previously describedfor neurogenin.

The p38 MAP kinase pathway and its implications in cellu-lar development. Numerous factors and ligands have beenidentified that promote neuronal and/or inhibit glial differen-tiation of neural stem cells; however, the intracellular signalingmechanisms and their effector molecules are still unknown formost of these factors. We suggest herein a novel mechanism bywhich the MAP kinase kinase kinase ASK1 induces neuronaldifferentiation in neural stem cells and, in parallel, effectivelyinhibits gliogenesis via activation of the p38 MAP kinase path-way. In a previous paper, Takeda et al. reported that inductionof survival and neuronal differentiation in undifferentiatedPC12 cells by ASK1 is dependent on activation of the p38MAP kinase pathway (39). However, progenitor cells of theperipheral and central nervous systems have quite differentproperties. Therefore, the p38 MAP kinase-mediated induc-tion of neuronal differentiation by ASK1 could be restricted tocells of the peripheral nervous system and might not necessar-ily be involved in neuronal differentiation of progenitor cellsderived from the CNS. Nonetheless, our data suggest that theASK1-p38-MAP kinase pathway is an important signalingpathway for neuronal differentiation of progenitor cells de-rived from the CNS. Furthermore, our results indicate thatastroglial development is effectively inhibited in adult neural

FIG. 9. ASK1-mediated inhibition of GFAP expression is depen-dent on p38 MAP kinase. (A) AHPs were transfected with hGFAP-Luc and infected with Ad-LacZ, Ad-ASK1-KM, or Ad-ASK1-�N 4 hafter transfection. Cells were treated with 10 �M SB 202580 or DMSOas a control at the time of infection, and luciferase activity was deter-mined after another 36 h. (B) GFAP promoter analysis on AHPstransfected with plasmids encoding either ASK1-KM or ASK1-�N aswell as with plasmids carrying pcDNA3, p38 wild type (p38wt), orp38DN. Luciferase activity was measured 36 h after transfection.(C) AHPs were incubated with 10 �M SB 203580 prior to infectionwith Ad-LacZ or Ad-ASK1-�N. SB 203580 was readded to the culturemedium 2 days after infection, and cells were harvested for Westernblot analysis for GFAP after another 2 days. The values representmeans � SEM (n � 3). An asterisk indicates a P value of 0.05,

and three asterisks indicate a P value of 0.001 in comparison to theLacZ-infected (A) or pcDNA-transfected (B) control; two numbersigns indicate a P value of 0.01, and three number signs indicate a Pvalue of 0.001 in comparison to the DMSO control infected with thesame viral construct (A) or to the KM-transfected control (B). Resultswere determined by ANOVA and the Scheffe post hoc test.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

progenitor cells when the p38 MAP kinase is activated follow-ing ASK1 signaling. It is reasonable to assume that, throughoutdevelopment, once a certain cellular fate is promoted, theopposing fate would be simultaneously inhibited. We suggestherein that ASK1 exerts via the p38 MAP kinase pathway adual role in development of immature cells of neural origin bypromoting neuronal fate on the one hand and by inhibitingastroglial fate on the other.

Possible implications for ASK1 in gene therapy of neuro-logical diseases. Stem cells are presently believed to be apromising cell source for cell replacement strategies in neuro-degenerative diseases such as Parkinson’s disease. However,due to poor survival rate and a low yield of cells adoptingneuronal phenotypes, neurotrophic factor delivery and geneticengineering of the cells prior to grafting are presently beinginvestigated in order to overcome these obstacles. Astroglialdifferentiation within the graft may be problematic, especiallywhen the cells are grafted into nonneurogenic areas such as thestriatum. Thus, tools preventing glial differentiation in thegrafted cells are of special interest. Our results indicate thatgenetic engineering of stem cells with ASK1 prior to transplan-tation might be a new strategy to improve the yield of devel-oping neurons. Since the ability of ASK1 to inhibit glial fateappears to be the major effect on AHPs, one might considercombining ASK1 with another gene or trophic factor that in-structively induces neuronal fate within the graft. Nonetheless,because a permanent ASK1 signaling in the transduced cells isnecessary at least for the inhibition of glial-cell development,we suggest the use of a system in which the expression levels ofASK1 remain stable, such as a retroviral infection system.

ACKNOWLEDGMENTS

We thank P. S. Eriksson for discussions and for reviewing the manu-script and J. E. Johnson, C.-H. Heldin, M. Landstrom, H. Ichijo, andF. H. Gage for the gifts of antibodies, cDNAs, and cells.

This work was supported by grants from the Swedish Medical Re-search Council, the Swedish Cancer Society, the Swedish Children’sCancer Society, the Hjalmar Svensson Forskningsfond, the Adler-bertska Forskningsstiftelsen, the Jubileum Clinic, and the MedicalFaculty at Goteborg University. H.U. was supported by the Wenner-Gren Foundation and the Wenner-Gren Society, the Uehara Memo-rial Life Science Foundation, and the Kanae Foundation for Life &Socio-Medical Science. A.B. was partly supported by the BoehringerIngelheim Foundation.

REFERENCES

1. Aberg, M. A., N. D. Aberg, H. Hedbacker, J. Oscarsson, and P. S. Eriksson.2000. Peripheral infusion of IGF-I selectively induces neurogenesis in theadult rat hippocampus. J. Neurosci. 20:2896–2903.

2. Aberg, M. A., F. Ryttsen, G. Hellgren, K. Lindell, L. E. Rosengren, A. J.MacLennan, B. Carlsson, O. Orwar, and P. S. Eriksson. 2001. Selectiveintroduction of antisense oligonucleotides into single adult CNS progenitorcells using electroporation demonstrates the requirement of STAT3 activa-tion for CNTF-induced gliogenesis. Mol. Cell. Neurosci. 17:426–443.

3. Altman, J., and G. D. Das. 1965. Autoradiographic and histological evidenceof postnatal hippocampal neurogenesis in rats. J. Comp. Neurol. 124:319–335.

4. Bonni, A., Y. Sun, M. Nadal-Vicens, A. Bhatt, D. A. Frank, I. Rozovsky, N.Stahl, G. D. Yancopoulos, and M. E. Greenberg. 1997. Regulation of glio-genesis in the central nervous system by the JAK-STAT signaling pathway.Science 278:477–483.

5. Brannvall, K., L. Korhonen, and D. Lindholm. 2002. Estrogen-receptor-dependent regulation of neural stem cell proliferation and differentiation.Mol. Cell. Neurosci. 21:512–520.

6. Brederlau, A., R. Faigle, P. Kaplan, P. Odin, and K. Funa. 2002. Bonemorphogenetic proteins but not growth differentiation factors induce dopa-minergic differentiation in mesencephalic precursors. Mol. Cell. Neurosci.21:367–378.

7. Casarosa, S., C. Fode, and F. Guillemot. 1999. Mash1 regulates neurogenesisin the ventral telencephalon. Development 126:525–534.

8. Cau, E., S. Casarosa, and F. Guillemot. 2002. Mash1 and Ngn1 controldistinct steps of determination and differentiation in the olfactory sensoryneuron lineage. Development 129:1871–1880.

9. Cau, E., G. Gradwohl, C. Fode, and F. Guillemot. 1997. Mash1 activates acascade of bHLH regulators in olfactory neuron progenitors. Development124:1611–1621.

10. Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA isolationby acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Bio-chem. 162:156–159.

11. Eriksson, P. S., E. Perfilieva, T. Bjork-Eriksson, A. M. Alborn, C. Nordborg,D. A. Peterson, and F. H. Gage. 1998. Neurogenesis in the adult humanhippocampus. Nat. Med. 4:1313–1317.

12. Farinas, I., M. Cano-Jaimez, E. Bellmunt, and M. Soriano. 2002. Regulationof neurogenesis by neurotrophins in developing spinal sensory ganglia. BrainRes. Bull. 57:809–816.

13. Gage, F. H., P. W. Coates, T. D. Palmer, H. G. Kuhn, L. J. Fisher, J. O.Suhonen, D. A. Peterson, S. T. Suhr, and J. Ray. 1995. Survival and differ-entiation of adult neuronal progenitor cells transplanted to the adult brain.Proc. Natl. Acad. Sci. USA 92:11879–11883.

14. Gross, R. E., M. F. Mehler, P. C. Mabie, Z. Zang, L. Santschi, and J. A.Kessler. 1996. Bone morphogenetic proteins promote astroglial lineage com-mitment by mammalian subventricular zone progenitor cells. Neuron 17:595–606.

15. Horton, S., A. Meredith, J. A. Richardson, and J. E. Johnson. 1999. Correctcoordination of neuronal differentiation events in ventral forebrain requiresthe bHLH factor MASH1. Mol. Cell. Neurosci. 14:355–369.

16. Ichijo, H., E. Nishida, K. Irie, P. ten Dijke, M. Saitoh, T. Moriguchi, M.Takagi, K. Matsumoto, K. Miyazono, and Y. Gotoh. 1997. Induction ofapoptosis by ASK1, a mammalian MAPKKK that activates SAPK/JNK andp38 signaling pathways. Science 275:90–94.

17. Johe, K. K., T. G. Hazel, T. Muller, M. M. Dugich-Djordjevic, and R. D.McKay. 1996. Single factors direct the differentiation of stem cells from thefetal and adult central nervous system. Genes Dev. 10:3129–3140.

18. Kanamoto, T., M. Mota, K. Takeda, L. L. Rubin, K. Miyazono, H. Ichijo, andC. E. Bazenet. 2000. Role of apoptosis signal-regulating kinase in regulationof the c-Jun N-terminal kinase pathway and apoptosis in sympathetic neu-rons. Mol. Cell. Biol. 20:196–204.

19. Kilpatrick, T. J., and P. F. Bartlett. 1995. Cloned multipotential precursorsfrom the mouse cerebrum require FGF-2, whereas glial restricted precursorsare stimulated with either FGF-2 or EGF. J. Neurosci. 15:3653–3661.

20. Lee, J. E., S. M. Hollenberg, L. Snider, D. L. Turner, N. Lipnick, and H.Weintraub. 1995. Conversion of Xenopus ectoderm into neurons byNeuroD, a basic helix-loop-helix protein. Science 268:836–844.

21. Lo, L. C., J. E. Johnson, C. W. Wuenschell, T. Saito, and D. J. Anderson.1991. Mammalian achaete-scute homolog 1 is transiently expressed by spa-tially restricted subsets of early neuroepithelial and neural crest cells. GenesDev. 5:1524–1537.

22. Mabie, P. C., M. F. Mehler, R. Marmur, A. Papavasiliou, Q. Song, and J. A.Kessler. 1997. Bone morphogenetic proteins induce astroglial differentiationof oligodendroglial-astroglial progenitor cells. J. Neurosci. 17:4112–4120.

23. Miyake, S., M. Makimura, Y. Kanegae, S. Harada, Y. Sato, K. Takamori, C.Tokuda, and I. Saito. 1996. Efficient generation of recombinant adenovirusesusing adenovirus DNA-terminal protein complex and a cosmid bearing thefull-length virus genome. Proc. Natl. Acad. Sci. USA 93:1320–1324.

24. Nakashima, K., T. Takizawa, W. Ochiai, M. Yanagisawa, T. Hisatsune, M.Nakafuku, K. Miyazono, T. Kishimoto, R. Kageyama, and T. Taga. 2001.BMP2-mediated alteration in the developmental pathway of fetal mousebrain cells from neurogenesis to astrocytogenesis. Proc. Natl. Acad. Sci. USA98:5868–5873.

25. Nakashima, K., M. Yanagisawa, H. Arakawa, and T. Taga. 1999. Astrocytedifferentiation mediated by LIF in cooperation with BMP2. FEBS Lett.457:43–46.

26. Palmer, T. D., J. Ray, and F. H. Gage. 1995. FGF-2-responsive neuronalprogenitors reside in proliferative and quiescent regions of the adult rodentbrain. Mol. Cell. Neurosci. 6:474–486.

27. Palmer, T. D., J. Takahashi, and F. H. Gage. 1997. The adult rat hippocam-pus contains primordial neural stem cells. Mol. Cell. Neurosci. 8:389–404.

28. Reynolds, B. A., and S. Weiss. 1996. Clonal and population analyses dem-onstrate that an EGF-responsive mammalian embryonic CNS precursor is astem cell. Dev. Biol. 175:1–13.

29. Richards, L. J., T. J. Kilpatrick, R. Dutton, S. S. Tan, D. P. Gearing, P. F.Bartlett, and M. Murphy. 1996. Leukaemia inhibitory factor or relatedfactors promote the differentiation of neuronal and astrocytic precursorswithin the developing murine spinal cord. Eur. J. Neurosci. 8:291–299.

30. Roy, N. S., S. Wang, L. Jiang, J. Kang, A. Benraiss, C. Harrison-Restelli,R. A. Fraser, W. T. Couldwell, A. Kawaguchi, H. Okano, M. Nedergaard, andS. A. Goldman. 2000. In vitro neurogenesis by progenitor cells isolated fromthe adult human hippocampus. Nat. Med. 6:271–277.

31. Saito, I., Y. Oya, K. Yamamoto, T. Yuasa, and H. Shimojo. 1985. Construc-


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

tion of nondefective adenovirus type 5 bearing a 2.8-kilobase hepatitis Bvirus DNA near the right end of its genome. J. Virol. 54:711–719.

32. Saitoh, M., H. Nishitoh, M. Fujii, K. Takeda, K. Tobiume, Y. Sawada, M.Kawabata, K. Miyazono, and H. Ichijo. 1998. Mammalian thioredoxin is adirect inhibitor of apoptosis signal-regulating kinase (ASK) 1. EMBO J.17:2596–2606.

33. Sayama, K., Y. Hanakawa, Y. Shirakata, K. Yamasaki, Y. Sawada, L. Sun, K.Yamanishi, H. Ichijo, and K. Hashimoto. 2001. Apoptosis signal-regulatingkinase 1 (ASK1) is an intracellular inducer of keratinocyte differentiation.J. Biol. Chem. 276:999–1004.

34. Scardigli, R., C. Schuurmans, G. Gradwohl, and F. Guillemot. 2001. Cross-regulation between Neurogenin2 and pathways specifying neuronal identityin the spinal cord. Neuron 31:203–217.

35. Suhonen, J. O., D. A. Peterson, J. Ray, and F. H. Gage. 1996. Differentiationof adult hippocampus-derived progenitors into olfactory neurons in vivo.Nature 383:624–627.

36. Sun, Y., M. Nadal-Vicens, S. Misono, M. Z. Lin, A. Zubiaga, X. Hua, G. Fan,and M. E. Greenberg. 2001. Neurogenin promotes neurogenesis and inhibitsglial differentiation by independent mechanisms. Cell 104:365–376.

37. Takahashi, J., T. D. Palmer, and F. H. Gage. 1999. Retinoic acid and

neurotrophins collaborate to regulate neurogenesis in adult-derived neuralstem cell cultures. J. Neurobiol. 38:65–81.

38. Takahashi, M., T. D. Palmer, J. Takahashi, and F. H. Gage. 1998. Wide-spread integration and survival of adult-derived neural progenitor cells in thedeveloping optic retina. Mol. Cell. Neurosci. 12:340–348.

39. Takeda, K., T. Hatai, T. S. Hamazaki, H. Nishitoh, M. Saitoh, and H. Ichijo.2000. Apoptosis signal-regulating kinase 1 (ASK1) induces neuronal differ-entiation and survival of PC12 cells. J. Biol. Chem. 275:9805–9813.

40. Tanigaki, K., F. Nogaki, J. Takahashi, K. Tashiro, H. Kurooka, and T.Honjo. 2001. Notch1 and Notch3 instructively restrict bFGF-responsive mul-tipotent neural progenitor cells to an astroglial fate. Neuron 29:45–55.

41. Verma-Kurvari, S., T. Savage, K. Gowan, and J. E. Johnson. 1996. Lineage-specific regulation of the neural differentiation gene MASH1. Dev Biol.180:605–617.

42. Young, M. J., J. Ray, S. J. Whiteley, H. Klassen, and F. H. Gage. 2000.Neuronal differentiation and morphological integration of hippocampal pro-genitor cells transplanted to the retina of immature and mature dystrophicrats. Mol. Cell. Neurosci. 16:197–205.

43. Zhu, G., M. F. Mehler, J. Zhao, S. Yu Yung, and J. A. Kessler. 1999. Sonichedgehog and BMP2 exert opposing actions on proliferation and differen-tiation of embryonic neural progenitor cells. Dev. Biol. 215:118–129.


Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/m

cb o

n 11

Nov

embe

r 20

21 b

y 17

8.15

8.25

0.88

.

ask1 inhibits astroglial development via p38 mitogen-activated

Documents