Enhanced autophagic cell death in expanded polyhistidine variants of HOXA1 reduces PBX1-coupled transcriptional activity and inhibits neuronal differentiation

Abstract

HOXA1 is a member of the homeobox gene family and is involved in early brain development. In our previous study, we identified novel variants of polyhistidine repeat tract in HOXA1 gene and showed that ectopic expression of expanded variants led to enhanced intranuclear aggregation and accelerated cell death in a time-dependent manner. Here, we further investigate the implications of polyhistidine variants on HOXA1 function. Aside from intranuclear aggregation, we observed cytosolic aggregates during the early stages of expression. Rapamycin, an autophagy inducer, resulted in decreased protein aggregation and cell death. Here, we also show an interaction between variants of HOXA1 and one of the HOX protein known cofactors, PBX1. Expanded HOXA1 variants exhibited reduced PBX1-coupled transcriptional activity through a regulatory enhancer of HOXB1. Moreover, we demonstrate that both deleted and expanded variants inhibited neurite outgrowth in retinoic acid-induced neuronal differentiation in neuroblastoma cells. These results provide further evidence that expanded polyhistidine repeats in HOXA1 enhance aggregation and cell death, resulting in impaired neuronal differentiation and cooperative binding with PBX1.

Full text

Enhanced Autophagic Cell Death in
Expanded Polyhistidine Variants of
HOXA1 Reduces PBX1-Coupled
Transcriptional Activity and Inhibits
Neuronal Differentiation
Rubigilda C. Paraguison,
1
Katsumi Higaki,
1
Kenji Yamamoto,
2
Hideo Matsumoto,
2
Tsukasa Sasaki,
3
Nobumasa Kato,
3
and Eiji Nanba
1
*
1
Division of Functional Genomics, Research Center for Bioscience and Technology, Tottori University,
Yonago, Japan
2
Department of Psychiatry and Behavioral Sciences, Tokai University School of Medicine,
Kanagawa, Japan
3
Department of Psychiatry, Graduate School of Medicine, University of Tokyo, Tokyo, Japan
HOXA1 is a member of the homeobox gene family and
is involved in early brain development. In our previous
study, we identified novel variants of polyhistidine re-
peat tract in HOXA1 gene and showed that ectopic
expression of expanded variants led to enhanced intra-
nuclear aggregation and accelerated cell death in a
time-dependent manner. Here, we further investigate the
implications of polyhistidine variants on HOXA1 func-
tion. Aside from intranuclear aggregation, we observed
cytosolic aggregates during the early stages of expres-
sion. Rapamycin, an autophagy inducer, resulted in de-
creased protein aggregation and cell death. Here, we
also show an interaction between variants of HOXA1
and one of the HOX protein known cofactors, PBX1.
Expanded HOXA1 variants exhibited reduced PBX1-
coupled transcriptional activity through a regulatory en-
hancer of HOXB1. Moreover, we demonstrate that both
deleted and expanded variants inhibited neurite out-
growth in retinoic acid-induced neuronal differentiation
in neuroblastoma cells. These results provide further
evidence that expanded polyhistidine repeats in HOXA1
enhance aggregation and cell death, resulting in im-
paired neuronal differentiation and cooperative binding
with PBX1. V
V
C2006 Wiley-Liss, Inc.
Key words: HOXA1; neuronal differentiation; PBX1; poly-
histidine; protein aggregation; autophagy
HOX genes form a subset of the family of home-
obox genes (Pearson et al., 2005). They are involved in
specifying positional identity along the anterior-posterior
axis of all bilaterian animals. In humans, the HOXA–D
clusters comprise 39 HOX genes, located on chromo-
some regions 7p15, 17p21, 12q13, and 2q31 (Grier et al.,
2005). During embryonic development, HOX genes
are expressed sequentially 30to 50along the anterior to
posterior axis. HOX genes contain a 61-amino-acid
helix-turn-helix DNA-binding domain known as the
homeodomain (Gehring et al., 1994). It is well established
that HOX/DNA binding specificity is modified by other
DNA-binding proteins, which act as cofactors. Among
these are the PBX proteins, which are widely expressed
in fetal and adult tissues and interact preferentially with
30HOX proteins (Phelan and Featherstone, 1997). PBX
can modulate the affinity and stability of DNA binding
and regulate transcriptional activity. The cooperative he-
terodimerization is carried out through a conserved
protein motif found N-terminal to the Hox homeodo-
main (Slupsky et al., 2001; Huang et al., 2005). Interac-
tions between PBX and HOX might also be mediated
by residues of the N-terminal arm of HOX proteins
(Shanmugam et al., 1997). It has been reported that
PBX1 and HOXB1 can cooperatively activate the tran-
scription through an autoregulatory element, directing
spatially restricted expression of the HOXB1 gene (b1-
ARE) in the developing hindbrain. However, only lim-
ited kinds of HOX can bind cooperatively with PBX
(HOXA1, HOXB1, and HOXA2; Di Rocco et al.,
Contract grant sponsor: Japanese Ministry of Health, Labor and Welfare;
Contract grant number: H14-kokoro-002; Contract grant sponsor: Japa-
nese Ministry of Education, Culture Sports, Science and Technology;
Contract grant number: 2005, 17659315; Contract grant number: 2004,
16012242.
*Correspondence to: Eiji Nanba, MD, PhD, Division of Functional
Genomics, Research Center for Bioscience and Technology, Tottori
University, 86 Nishi-machi, Yonago 683-8503, Japan.
E-mail: enanba@grape.med.tottori-u.ac.jp
Received 18 April 2006; Revised 4 October 2006; Accepted 5 October
2006
Published online 27 November 2006 in Wiley InterScience (www.
interscience.wiley.com). DOI: 10.1002/jnr.21137
Journal of Neuroscience Research 85:479–487 (2007)
'2006 Wiley-Liss, Inc.

2001). This autoregulatory enhancer is the key regula-
tory element for the normal rhombomere 4 expression
of Hoxb1 in the developing hindbrain, whereas Hoxa1
and Hoxb1 synergize in patterning the hindbrain cranial
nerves and second pharyngeal arch (Gavalas et al., 1998).
HOXA1 is one of the first HOX genes to be
expressed during embryonic development (Pearson et al.,
2005). In mice, its expression starts from 7.5 dpc, and it
is established in the neuroectoderm and mesoderm at
8.0 dpc. (Remacle et al., 2004). HOXA1 gene encodes
two alternatively spliced mRNAs, which appear to be
differentially expressed in the developing embryo. The
homeodomain-containing variant 1 undergoes transcrip-
tional activation during hindbrain development from E7
to E8.5 (Godwin et al., 1998). Functional inactivation of
this gene results in prenatal lethality and numerous mal-
formations (Lufkin et al., 1991; Carpenter et al., 1993).
Hoxa1 null mice exhibit hindbrain segmentation and pa-
tterning defects that cause abnormal development of
cranial nerve, cranial ganglia, and branchial arch deriva-
tives (Chisaka et al., 1992). Furthermore, ectopic expres-
sion of Hoxa1 in transgenic mice leads to abnormalities
of the developing hindbrain and ultimately results in
embryonic death (Zhang et al., 1994).
HOXA1 gene contains a tract of 10-histidine re-
peat. In our previous study, we identified novel variants
of polyhistidine tracts in HOXA1 gene in a Japanese
population, and no homozygous case has been found for
any of these variants (Paraguison et al., 2005). Certain
individuals were heterozygous for deleted 7- and 9-histi-
dine repeats and expanded 11- and 12-histidine repeats.
Expression of expanded polyhistidine variants of HOXA1
proteins resulted in accelerated formation of ubiquitinated
intranuclear aggregates and increased cell death. How-
ever, the mechanism by which this aggregation occurs is
poorly understood. In this study, we showed that expres-
sion of expanded polyhistidine variants of HOXA1 in
human neuroblastoma cell line SK-N-SH and embryonic
carcinoma cell line P19 also caused increased intranuclear
aggregation and cell death. However, there was a signifi-
cant reduction of protein aggregation and cell death
upon Rapamycin treatment, indicating involvement of
the autophagic process. Expanded variants exhibited im-
paired cooperative binding with the cofactor PBX1,
resulting in decreased transcriptional activity. Moreover,
cells overexpressing expanded and deleted variants ex-
hibited impaired neuronal differentiation. These data
provide new insights on the function of polyhistidine
variants of HOXA1 protein in neuronal cells.
MATERIALS AND METHODS
Antibodies and Reagents
The following antibodies were used: polyclonal goat
anti-HOXA1 antibody (Santa Cruz Biotechnology, Santa
Cruz, CA; sc-17146), monoclonal mouse anti-EGFP antibody
(Santa Cruz Biotechnology, sc-9996), polyclonal rabbit anti-
MAP2, H-300 (Santa Cruz Biotechnology, sc-20172), poly-
clonal rabbit anti PBX1, P-20 (Santa Cruz Biotechnology,
sc-889), and polyclonal rabbit anti-b-tubulin H235 (Santa
Cruz Biotechnology, sc-9104) and Alexa Fluor 594-conjugated
anti-goat IgG, Alexa Fluor 555-conjugated anti-rabbit IgG,
and Alexa Fluor 555-conjugated anti-mouse IgG (Molecular
Probes, Eugene, OR). The following reagents were also used
for this study: 10 lM retinoic acid (Sigma, St. Louis, MO;
R2625) immediately after transfection, 10 mM 3-methyladenine
(3-MA; Sigma, M9281) 15 hr prior to fixation, and 100 lM
z-VAD-fmk (Promega, Madison, WI; G7231) and 2 lg/ml
Rapamycin (Sigma, R0395) both right after transfection.
Expression Vectors and Reporter Construct
Human HOXA1 expression vectors were generated as
described previously (Paraguison et al., 2005). The following
primer sets with suitable restriction enzyme recognition sites
(Sac I and Bam HI) were used to generate HOXA1 variant 1:
50- TAGAGCTCACCATGGACAATGCAAGAATGAACT-
CC-30and 50-ATGGATCCGTGTGGGAGGTAGTCAGA-
GTGTCTGA-30. All DNA amplification steps were per-
formed using high fidelity Pfu Ultra DNA polymerase (Strata-
gene, La Jolla, CA) using genomic DNA and cDNA derived
from human normal lymphoblast and confirmed by sequenc-
ing (ABI 3130xl). Expression constructs were derived from
CMV promoter-based expression vector, pCMV-Script (Stra-
tagene) and pEGFP-N1 (Invitrogen, San Diego, CA). The
expression plasmids pCMV-Script and pEGFP-N1 were used
as empty vector controls. The luciferase reporter construct
pAdMLARE containing b1-ARE and the Pbx1 expression
construct pSGPbx1a were generous gifts of Prof. Zappavigna
(Di Rocco et al., 2001).
Cell Culture and Transfection
COS-7 cells, murine P19 embryonal carcinoma (EC)
cells, and the human neuroblastoma cell line SK-N-SH were
maintained in Dulbecco’s modified Eagle’s medium (DMEM;
Sigma; D6429) supplemented with 10% fetal bovine serum
(FBS; HyClone, Logan, UT). Cultures 50–80% confluent
were transfected with Fugene 6 transfection reagent (Roche,
Indianapolis, IN) in accordance with the manufacturer’s rec-
ommendations. For a typical experiment, 1 lg of expression
vector was used in a 35-mm culture dish. For cotransfection
experiments, 1 lg of reporter plasmid (pAdMLARE), 0.5 lg
of HOXA1 expression construct, 1 lg of PBX1 expression
construct, and 0.3 lg of pEGFP as an internal control were
used in a 35-mm dish. For Western blot analyses, cells cultured
in 10-cm dishes were transfected with 10 lg of plasmid con-
struct using Fugene-6 (Roche). To initiate differentiation,
SK-N-SH cells were inoculated and treated with 10 lM
retinoic acid (RA), whereas P19 cells were grown in a serum-
free condition. Dead cells were scored under a fluorescence
microscope (Leica DMIRE2).
Luciferase Assay
After 24 hr posttransfection, the cells were harvested
and lysed in Pica Gene cell culture lysis reagent (Toyo Ink,
Tokyo, Japan). Luciferase assay was carried out by using a
Pica Gene kit (Toyo Ink) in accordance with the manufac-
turer’s protocol. HOXA1 cooperative expression with PBX1
480 Paraguison et al.
Journal of Neuroscience Research DOI 10.1002/jnr

using pAdMLARE reporter construct was assessed by quanti-
tative luciferase assay with Luminescencer-PSN (Bio-Instru-
ment ATTO AB-220). pEGFP-N1 plasmid was used as an in-
ternal control.
Immunostaining, Dead Cell Scoring, and Imaging
The cells transiently expressing HOXA1 constructs were
grown in 35-mm dishes containing glass coverslips. Cells
attached to the glass coverslips were washed with phosphate-
buffered saline (PBS), fixed with 4% paraformaldehyde, and
incubated with antibodies as described previously (Paraguison
et al., 2005). Confocal scanning analysis was performed with a
Leica confocal microscope (Leica TCS-SP2). The degree of
protein accumulation within the cell nuclei and the cytosol of
transfected cells was scored. The cells were counted from
10 randomly selected microscope fields of each sample, 15 in
the case of SK-N-SH cells. The ratio of the number of cells
with protein aggregation over the total number of cells was
then computed. Each experiment was performed independ-
ently in triplicate.
Western Blotting and Immunoprecipitation
Posttransfected cells were harvested and lysed by sonica-
tion in a buffer containing 10 mM Tris-HCl (pH 7.4),
150 mM NaCl, 1 mM EDTA, 1 mM EGTA, and protease
inhibitor cocktail (Roche). Proteins were quantified by using
a protein assay rapid kit (Wako, Osaka, Japan), run on 10% or
12% SDS-PAGE gels, and transferred on PVDF membranes
(Millipore, Bedford, MA; IPVH00010) by a semidry blotter
(Bio-Rad, Hercules, CA). Membranes were incubated with
antibodies as described previously. Protein lysates from cotrans-
fection experiments were immunoprecipitated with anti-GFP
and used for Western blotting with anti-PBX1. Polyclonal
rabbit anti-MAP2 antibody was used for quantifying neuronal
differentiation in SK-N-SH. Signals were detected using ECL
reagent (Amersham-Pharmacia Biotech, Arlington Heights, IL)
on X-ray films (Fuji). For quantification, images were ana-
lyzed in NIH Image software.
RESULTS
Expanded Polyhistidine Variants of HOXA1
Enhanced Intranuclear Protein Aggregation
in Neuronal Cell Lines
We have previously reported that expanded poly-
histidine variants in HOXA1 resulted in early nuclear
protein aggregation and an increased cell death in COS7
cells (Paraguison et al., 2005). Increasing polyhistidine
repeat length coincided with early protein aggregation
and accelerated cell death. In contrast, no significant dif-
ference was observed between cells overexpressing 7-poly-
histidine variant and wild-type 10-polyhistidine repeat
variant.
In the current study, we observed that these nu-
clear aggregations were also detected in both SK-N-SH
cells (Fig. 1A) and P19 cells (Fig. 1B). To rule out the
possibility that these aggregates were caused by enhanced
green fluorescent protein (EGFP) tagging, transfection
with untagged cytomegalovirus (CMV)-driven expres-
sion of HOXA1 variants was performed in COS-7 cells.
Immunostaining with anti-HOXA1/Alexa 594 anti-goat
also exhibited protein aggregates 18 hr posttransfection
(Fig. 1C). After 18–20 hr, cells transfected with EGFP-
tagged HOXA1 proteins were analyzed for Western
blotting. In lysates of cells transfected with 11- and 12-
polyhistidine variants, insoluble high-molecular-weight
proteins remained in the stacking gel (Fig. 1D), indica-
tive of protein complex accumulation.
Rapamycin Cleared Protein Aggregations and
Decreased Cell Death in COS-7 Cells Expressing
Expanded Variants of HOXA1, Whereas 3-MA
Reversed This Effect
We also attempted to determine the type of cell
death occurring as a consequence of expression of
HOXA1 expanded polyhistidine variants. Recently, it
has been reported that autophagic cell death was partially
mediated by caspase activation (Ravikumar et al., 2006).
Thus, we examined whether caspase inhibition could
suppress cell death by using a cell-permeable pan caspase
inhibitor, z-VAD-fmk, that binds to the catalytic site of
caspase proteases and can inhibit induction of apoptosis
(Broustas et al., 2004). However, no significant inhibi-
tion was seen in cells treated with 100 lM z-VAD-fmk
(Fig. 2A,B). Thus, we speculated that classical apoptosis
is not the mechanism involved in HOXA1-related cell
death.
To understand better the mechanism responsible
for clearance of protein aggregates, we examined the
possible role of autophagy in degrading these proteins and
the effect of caspase inhibition on cell death in COS-7
cells. During our time course experiments, we observed
that aggregations were present not only in the nucleus
but also in the cytosol at the early stages of expression
(18 hr after transfection). These cytosolic aggregates
were abundant in cells expressing the expanded 12-histi-
dine variants of HOXA1-EGFP (Fig. 2C). Aggregates
gradually cleared out by the endogenous autophagic
mechanism of the cell and were concentrated in the
nuclei at 42 hr after transfection (Fig. 2C,D). We ex-
plored the possible involvement of the autophagic pro-
cess in the clearing of protein aggregates in these cells.
3-Methyladenine (3-MA), which has an inhibitory ef-
fect on autophagy (Ravikumar et al., 2002), enhanced
nuclear aggregations and cell death (Fig. 2E,F). There is
a decreased rate of clearance of protein aggregates
observed in 3-MA-treated cells overexpressing 12-His
variant. Cytosolic aggregates around the outer perinu-
clear periphery are noticeable in the later stage of expres-
sion, i.e., about 42 hr, where they were supposed to
be cleared out in the 3-MA nontreated transfected cells
(Fig. 2F). Conversely, rapamycin, an autophagy-inducing
chemical (Ravikumar et al., 2006), decreased cytosolic and
nuclear aggregation that eventually reduced cell death
particularly in the expanded 12-histidine variant (Fig. 2G).
This suggested that clearing of HOXA1-EGFP protein
aggregates was mediated by an autophagic mechanism. We
Polyhistidine Variants of HOXA1 in Neurons 481
Journal of Neuroscience Research DOI 10.1002/jnr

further noted that the autophagic clearing system was
more efficient in degrading cytosolic aggregates, since
autophagic vacuoles were directly accessible in the cytosol.
However, the increased production of accumulated intra-
nuclear proteins might have overloaded the autophagic
clearing mechanism, thus resulting in cell death.
Fig. 1. Expression of HOXA1 polyhistidine expansion variants results
in protein aggregation. HOXA1-EGFP construct variants were expressed
in neuronal cell lines. As early as 15–20 hr after transfection, intranuclear
aggregates were already detected in cells transfected with expanded
HOXA1 variant. Images were taken 18 hr after transfection. EGFP fluo-
rescent signals were visualized by confocal microscopy. A: Neuroblas-
toma cells, SK-N-SH. B: Embryonic carcinoma cells, P19. C: COS-7
cells transfected with untagged HOXA1 and immunostained with anti-
HOXA1/Alexa 594 anti-goat (red signals). Fluorescence images were
taken 18 hr posttransfection. D: 18–24-hr posttransfected cell lysates an-
alyzed for Western blotting using anti-GFP. An increased amount of
high-molecular-weight SDS-insoluble protein was detected in the stack-
ing gel in lanes of expanded HOXA1, indicating that the length of histi-
dine repeats is directly proportional with the degree protein accumula-
tions. [Color figure can be viewed in the online issue, which is available
at www.interscience.wiley.com.]
Fig. 2. Protein aggregations in the cytosol are cleared by an autopha-
gic process. A: Right after transfection, cells were treated with
100 lM z-VAD-fmk and scored after 24 hr postransfection. z-VAD-
fmk did not inhibit intranuclear protein aggregation. Mock EGFP
served as a negative control. B: Fluorescence images of EGFP and
HOXA1-EGFP 12-His variant treated with z-VAD-fmk. C: COS-7
cells were transfected with HOXA1-EGFP constructs at 18 hr post-
transfection; cytosolic aggregates could be seen abundantly in cells
transfected with HOXA1-EGFP 12-histidine variant. These aggre-
gates were eventually cleared out after 42 hr and were mostly con-
centrated in the nuclei. D: Graph showing percentage of cells with
protein aggregation and dead cells per number of EGFP-positive
transfected cells at 18 and 42 hr after transfection. E: Transfected cells
were treated with 10 mM 3-methyladenine (3-MA) 15 hr prior to
fixation. HOXA1-EGFP 10-His variant-transfected cells were treated
with DMSO to serve as a control for cell toxicity. 3-MA increases
protein aggregations and cell death in HOXA1-EGFP-transfected
COS-7 cells. Dead cells and those exhibiting nuclear and cytosolic
aggregation were scored 18 hr after transfection. F: No protein
aggregation was detected in mock-EGFP-transfected COS-7 cells
treated with 3-MA; however, an increase in cell death was observed.
HOXA1-EGFP 12-His variant exhibited cytosolic aggregates around
the outer perinuclear periphery, which are denoted by arrows. G:
Immediately after transfection, 10-His and 12-His repeat variants of
HOXA1-EGFP-transfected COS-7 cells were treated with 2 lg/ml
Rapamycin prior to fixation. Rapamycin reduces cytosolic aggrega-
tion and cell death significantly in the cells transfected with 12-His
repeat variant. Dead cells and protein aggregations were scored 42 hr
after transfection. Error bars represent SEM; n ¼3. *P<0.05,
**P<0.01. Pvalues from a paired t-test in all experiments. [Color
figure can be viewed in the online issue, which is available at
www.interscience.wiley.com.]
"
482 Paraguison et al.
Journal of Neuroscience Research DOI 10.1002/jnr

Figure 2.
Polyhistidine Variants of HOXA1 in Neurons 483
Journal of Neuroscience Research DOI 10.1002/jnr

Transcriptional Activities of Polyhistidine
Variants of HOXA1 Coupled With PBX1
We next examined the physiological relevance of
polyhistidine variants of HOXA1 protein in neuronal
cells. PBX1 plays a major role in cooperative transcrip-
tional activation with HOXA1 through an autoregula-
tory element, the b1-ARE (Di Rocco et al., 2001). To
investigate the cooperative transcription of PBX1 with
Fig. 3. Expanded HOXA1 reduces PBX1-coupled transcriptional
activity. Luciferase assay from the different cell lines. Cells were
cotransfected with plasmids containing PBX1, HOXA1, and the
enhancer b1-ARE fused to a luciferase reporter gene (Di Rocco et al.,
2001). EGFP vector was used as an internal control. A: COS-7. B:
Neuroblastoma cell line SK-N-SH. C: Embryonic carcinoma cell line
P19 grown in serum-free medium. Bars represent the mean 6SE of
at least three independent experiments. *P<0.05, **P<0.01.
Fig. 4. HOXA1 and PBX1 interaction is impaired in expanded
variant of HOXA1. A: Immunofluorescence images of COS-7
cotransfected with PBX1, an enhancer bi-ARE, and 7-His, 10-His, or
12-His repeat variants of HOXA1-EGFP. Fluorescence images were
taken 24 hr posttransfection. Red signals: anti-PBX1. Note the absence
of PBX1 signal in protein aggregates of the 12-His variant. B: Immu-
noprecipitation with anti-GFP and Western blotted with anti-PBX1
and anti-HOXA1. PBX1 protein shows decreased levels in both
expanded and deleted variants of HOXA1-EGFP transfected to COS-
7cells.C: Quantification of bound PBX1 protein levels per HOXA1
protein from Western blot analysis. [Color figure can be viewed in the
online issue, which is available at www.interscience.wiley.com.]
484 Paraguison et al.
Journal of Neuroscience Research DOI 10.1002/jnr

the different variants of HOXA1 through b1-ARE, we
simultaneously expressed HOXA1, PBX1, EGFP (an
internal control), and b1-ARE luciferase reporter con-
structs in COS7, SK-N-SH, and P19 cells. Luciferase
assay were performed 24 hr after transfection. The P19
embryonic cell line was grown under serum-free condi-
tions. Growth in serum-free media itself committed EC
cells to neural differentiation (Darmon et al., 1981) and
addition of retinoic acid (RA) intensified this effect
(Tanaka et al., 1992). However, nonneural cell types arise
after treatment with RA under serum-containing condi-
tions. Differentiation in serum-free media alone is accom-
panied by expression of only neuroectodermal/neural
mRNAs, but treatment with RA invariably induces the
cells to express both neuroectodermal/neural and endo-
dermal mRNAs (Pachernik et al., 2005), so the serum-
free medium condition was used in this experiment.
The transcriptional activities of expanded variants
were significantly reduced in all cell lines, whereas the
activities of deleted variant with 7-polyhistidine were
enhanced in P19 cells compared with the activity of
10-polyhistidine variant (Fig. 3A–C). Immunofluores-
cence images revealed the inability of aggregated forms
of expanded HOXA1 protein to bind efficiently with
PBX1 protein (Fig. 4A). Inefficient interaction between
expanded HOXA1 and PBX1 is clearly indicated by
low levels of PBX1 protein in immunoprecipitation assay
with anti-GFP (Fig. 4B,C).
Expanded and Deleted Polyhistidine Variants in
HOXA1 Inhibited Neuronal Differentiation
Because HOXA1 is one of the neurodevelopmental
genes, we examined whether polyhistidine variants had
any effect on the process of neuronal differentiation. To
accomplish this, RA-induced neuronal differentiation
was performed in SK-N-SH cells. Phase microscopy
shows the effect of RA treatment on the neurites of these
cells (Fig. 5A). Immunocytochemistry with antibody
against the neuron-specific protein MAP2 revealed
neurite outgrowth in cells expressing the 10-histidine
HOXA1-GFP, 42 hr after transfection (Fig. 5B). In con-
trast, a greater number of cells expressing expanded and
deleted forms of HOXA1-GFP failed to initiate neurite
outgrowth. Total levels of MAP2 protein in cells over-
expressing expanded and deleted variants were also
significantly reduced (Fig. 5C,D). The appearance of
Fig. 5. Expanded and deleted polyhistidine repeats in HOXA1 in-
hibit neuronal differentiation. A: Retinoic acid (RA)-induced neuro-
nal differentiation in SK-N-SH cells Phase images of SK-N-SH cells
treated with (right) or without (left) 10 lM RA for 42 hr. B:
HOXA1-EGFP-transfected SK-N-SH treated with 10 lM RA for
42 hr. Immunofluorescence images where taken 42 hr after transfec-
tion. Expanded 11- and 12-His repeat variants as well as deleted
7-His repeat variants inhibit neuronal differentiation. Note the neurite
outgrowth immunostained with MAP2 (red signal). C: Western blot
analysis shows decreased MAP2 expression levels in the expanded and
deleted forms. D: Quantification of MAP2 expression levels normal-
ized with b-tubulin. E: Quantification of the percentage of EGFP-
MAP2-double-positive cells. Error bars represent SEM; n ¼3. *P<
0.05, **P<0.01. Pvalues from a paired t-test in all experiments.
[Color figure can be viewed in the online issue, which is available at
www.interscience.wiley.com.]
"
Polyhistidine Variants of HOXA1 in Neurons 485
Journal of Neuroscience Research DOI 10.1002/jnr

differentiated cells was also examined by scoring MAP2-
positive transformants expressing the HOXA1-EGFP
variants. EGFP-MAP2 double-positive cells quantifica-
tion also confirmed that there is a significant decrease
in the percentage of MAP2 levels particularly in the
expanded variants.
DISCUSSION
In this current study, we have shown that poly-
histidine repeat tract variants compromised HOXA1
function in transcription and neuronal differentiation.
Moreover, we have presented evidence that cell death
resulting from polyhistidine repeat expansion is mediated
primarily by authophagy and not by a caspase-dependent
mechanism. Authophagy is a type of cell death involving
bulk degradation of cytoplasmic proteins or organelles in
the lytic compartment. Inhibition of autophagy enhances
protein aggregation and cell death (Ravikumar et al.,
2002). Our data showed that 3-MA accelerates protein
accumulations and enhances cell death, particularly in
the expanded variants (Fig. 2E,F). Conversely, Rapamy-
cin, an inducer of the autophagic process, decreases pro-
tein aggregates and reduces cell death (Fig. 2G). Taken
together, our findings indicate that autophagy is involved
in the degradation and clearance of aggregations of
expanded HOXA1 variants.
HOXA1 splice variant 1 is reported to be active in
E7–E8.5 and functions in the presumptive rhombomere
3 and 4 regions of the developing hindbrain (Zhang
et al., 1994). However, the nonhomeodomain-contain-
ing variant is expressed in the endodermal derivative
after E8.5 to the adult stage (Godwin et al., 1998).
Therefore, we speculate that splice variant 1 may be
active only in early stages of embryonic neurodevelop-
ment. As one of the DNA-binding proteins and tran-
scription factors in the HOX gene family, HOXA1 also
relies on the activity of transcriptional cofactors aside
from its DNA-binding properties (Pearson et al., 2005).
A previous study has reported that HOXB1 and PBX1
cooperatively activate transcription under the control of
b1-ARE (autoregulatory enhancer). Moreover, HOXA1
and HOXA2 are also able to activate transcription by
b1-ARE in cooperation with PBX1 (Di Rocco et al.,
2001). We examined the binding ability of our HOXA1
variants to PBX1 and b1-ARE by evaluating the coop-
erative transcriptional activation using a luciferase assay.
Expanded variants transfected into COS-7 as well as SK-
N-SH neuroblastoma and P19 embryonic carcinoma
cells (EC) that were induced to undergo neuronal differ-
entiation indeed showed significantly reduced activation
compared with the wild type, whereas the deleted
variant expressed in P19 cell line showed an enhanced
activation (Fig. 3). However, the reason for increased
transcriptional activity observed in the deleted variant in
this cell line is as yet unknown. Immunofluorescence
images and immunoprecipitation assay clearly confirmed
the inability of aggregated forms of expanded HOXA1
protein to bind efficiently with PBX1 protein. The co-
operative interaction between Hox and Pbx is mediated
by a conserved hexapeptide sequence located toward the
N-terminal region from Hox homeodomain (Phelan and
Fetherstone, 1997; Remacle et al., 2004). Our results
suggest that polyhistidine variants in HOXA1 might also
affect the binding efficiency of its homeodomain to
certain cofactors and/or other target genes or proteins.
Human neuroblastoma SK-N-SH cells, from a ma-
lignant pediatric tumor derived from the neural crest,
retains its ability to differentiate into the neuronal lineage
when exposed to RA (Wainwright et al., 2001).
HOXA1 is the first target gene activated by RA, fol-
lowed by a sequential activation of other HOX genes
(Simeone et al., 1990; Martinez-Ceballos et al., 2005).
We showed an inhibition of neuronal differentiation
by not only the extended but also the deleted variants
(Fig. 5). Our study provides novel insights on the patho-
logical implications of the polyhistidine tract in HOXA1
and leaves us with the intriguing possibility that polyhis-
tidine repeat expansions and deletions may cause aberra-
tions in neuronal morphogenesis or differentiation in
general.
EC cells differentiate into various lineages depend-
ing on the presence of activators in the culture medium
(Pachernik et al., 2005). Growth under serum-free con-
ditions committed EC cells to neural differentiation.
Differentiation of EC cells into endodermal-like cells is
induced by serum. Interestingly, we found that, in EC
cells committed to neural differentiation, transcriptional
activity inversely coincided with polyhistidine repeat
length. These results further strengthen our theory that
polyhistidine length affects neurodevelopment. In agree-
ment with our results, another study has proposed that
HOXA1 may function as a stimulator of neuroectoder-
mal and mesodermal differentiation and a repressor
of embryonic endoderm formation (Martinez-Ceballos
et al., 2005) and that aberration in HOXA1 could lead
to increased expression of endodermal genes by RA
and would lead to repression of neuroectodermal and
mesodermal markers. Premature death of HOXA1-
expressing cells may impair transcription and neuronal
differentiation.
Recently, Tischfield and colleagues (2005) reported
that patients with Bosley-Salih-Alohrainy syndrome were
homozygous for HOXA1 truncating mutations, whereas
heterozygotes had normal phenotypes. This mutation
resulted in abnormal development of the central nervous
system in the brainstem. Distinguishing phenotypes
include horizontal gaze abnormalities, mental retardation,
and autism spectrum disorder. Even low levels of Hoxa1
expression in Hoxa1
+/–
cells are sufficient for normal
activation of the Hoxa1 pathway and may explain why
mice heterozygous for HOXA1 mutations appear normal
(Pasqualetti et al., 2001; Martinez-Ceballos et al., 2005).
In our previous report, we identified heterozygous poly-
histidine repeat variants in HOXA1 gene from a
Japanese population comprising normal and autistic indi-
viduals. No individuals homozygous for these mutations
have been found. We speculate that there is a possibility
486 Paraguison et al.
Journal of Neuroscience Research DOI 10.1002/jnr

that phenotypic aberrations may exist in homozygous
individuals. Future in vivo studies are essential to exam-
ine the physiological functions of the polyhistidine var-
iants of HOXA1.
ACKNOWLEDGMENTS
We thank Prof. Vincenzo Zappavigna for provid-
ing the PBX1 and pAdMLARE reporter plasmids.
R.C.P. is the recipient of a research scholarship from the
Japanese Ministry of Education, Culture, Sports, Science
and Technology.
REFERENCES
Broustas CG, Gokhale PC, Rahman A, Dritschlo A, Ahmad I, Kasid U.
2004. BRCC2, a novel BH3-like domain-containing protein, induces
apoptosis in a caspase-dependent manner. J Biol Chem 279:26780–
26788.
Carpenter EM, Goddard JM, Chisaka O, Manley NR, Capecchi MR.
1993. Loss of Hoxa1 (Hox-1.6) function results in the reorganization of
the murine hindbrain. Development 118:1063–1075.
Chisaka O, Musci TS, Cappecchi MR. 1992. Developmental defects of
the ear, cranial nerves and hindbrain resulting from targeted disruption
of the mouse homeobox gene, Hox-1.6. Nature 355:516–520.
Darmon M, Bottenstein J, Sato G. 1981. Neural differentiation following
culture of embryonal carcinoma cells in a serum-free defined medium.
Dev Biol 85:463–473.
Di Rocco G, Gavalas A, Popperl H, Krumlauf R, Mavilio F, Zappavigna V.
2001. The recruitment of SOX/OCT complexes and the differential acti-
vity of HOXA1 and HOXB1 modulate the Hoxb1 auto-regulatory
enhancer function. J Biol Chem 276:20506–20515.
Gavalas A, Studer M, Lumsden A, Rijli FM, Krumlauf R, Chambon P.
1998. Hoxa1 and Hoxb1 synergize in patterning the hindbrain, cranial
nerves and second pharyngeal arch. Development 125:1123–1136.
Gehring WJ, Affolter M, Burglin T. 1994. Homeodomain proteins.
Annu Rev Biochem 63:487–526.
Godwin AR, Stadler HS, Nakamura K, Cappecchi MR. 1998. Detection
of targeted GFP-Hox gene fusions during mouse embryogenesis. Proc
Natl Acad Sci U S A 95:13042–13047.
Grier DG, Thompson A, Kwasniewska A, McGonigle GJ, Halliday HL,
Lappin TR. 2005. The pathophysiology of HOX genes and their role
in cancer. J Pathol 205:154–171.
Huang H, Rastegar M, Bodner C, Goh SL, Rambaldi I, Featherstone M.
2005. MEIS C termini harbor transcriptional activation domains that
respond to cell signaling. J Biol Chem 280:10119–10127.
Lufkin T, Dierich A, LeMeur M, Chambon M. 1991. Disruption of the
Hox-1.6 homeobox gene results in defects in a region corresponding to
its rostral domain of expression. Cell 66:1105–1119.
Martinez-Ceballos E, Chambon P, Gudas LJ. 2005. Differences in gene
expression between wild type and Hoxa1 knockout embryonic stem
cells after retinoic acid treatment or LIF removal. J Biol Chem 280:
16484–16498.
Pachernik J, Bryja V, Esner M, Kubala L, Dvorak P, Hampl A. 2005.
Neural differentiation of pluripotent mouse embryonic carninoma cells
by retinoic acid: inhibitory effects of serum. Physiol J 54:115–122.
Paraguison RC, Higaki K, Sakamoto Y, Hashimoto O, Miyake N,
Matsumoto H, Yamamoto K, Sasaki T, Kato N, Nanba E. 2005. Poly-
histidine tract expansions in HOXA1 result in intranuclear aggregation
and increased cell death. Biochem Biophys Res Commun 336:1033–
1039.
Pasqualetti M, Neun R, Davenne M, Rijli FM. 2001. Retinoic acid
rescues inner ear defects in Hoxa1 deficient mice. Nat Genet 1:34–39.
Pearson JC, Lemons D, McGinnis W. 2005. Modulating hox gene func-
tions during animal body patterning. Nat Rev Neurosci 6:893–904.
Phelan ML, Featherstone MS. 1997. Distinct HOX N-terminal arm
residues are responsible for specificity of DNA recognition by HOX
monomers and HOX-PBX heteromers. J Biol Chem 272:8635–8643.
Ravikumar B, Duden R, Rubinsztein DC. 2002. Aggregate-prone
proteins with polyglutamine and polyalanine expansions are degraded
by autophagy. Hum Mol Genet 11:1107–1117.
Ravikumar B, Berger Z, Vacher C, O’Kane CJ, Rubinszten DC. 2006.
Rapamycin pre-treatment protects against apoptosis. Hum Mol Genet
15:1209–1216.
Remacle S, Abbas L, De Backer O, Pacico N, Gavalas A, Gofflot F,
Picard JJRezsohazy R. 2004. Loss of function but no gain of function
caused by amino acid substitutions in the hexapeptide of Hoxa1 in vivo.
Mol Cell Biol 24:8567–8575.
Shanmugam K, Featherstone MS, Saragovi HU. 1997. Residues flanking
the HOX YPWM motif contribute to cooperative interactions with
PBX. J Biol Chem 272:19081–19087.
Simeone A, Acampora D, Arcioni L, Andrews PW, Boncinelli E,
Mavilio F. 1990. Sequential activation of HOX 2 homeobox genes by reti-
noic acid in human embryonal carcinoma cells. Nature 346:763–766.
Slupsky CM, Sykes DB, Gay GL, Sykes BD. 2001. The HoxB1 hexa-
peptide is a prefolded domain: implications for the Pbx1/Hox interac-
tion. Prot Sci 10:1244–1253.
Tanaka Y, Kawahata K, Nakata T, Hirokawa A. 1992. Chronological
expression of microtubule-associated proteins (MAPs) in EC cell P19
after neural induction by retinoic acid. Brain Res 596:269–278.
Tischfield MA, Bosley TM, Salih MAM, Alorainy IA, Sener EC, Nester
MJ, Oystreck DT, Chan WM, Andrews C, Erickson RP, Engle EC.
2005. Homozygous HOXA1 mutations disrupt human brainstem,
inner ear, cardiovascular and cognitive development. Nat Genet 37:
1035–1037.
Wainwright LJ, Lasorella A, Iavarone A. 2001. Distinct mechanisms of
cell cycle arrest control the decision between differentiation and senes-
cence in human neuroblastoma cells. Proc Natl Acad Sci U S A 98:
9396–9400.
Zhang M, Kim HJ, Marshall H, Gendron-Maguire M, Lucas DA, Baron A,
Gudas T, Gridley LJ, Krumlauf R, Grippo JF. 1994. Ectopic Hoxa-1
induces rhombomere transformation in mouse hindbrain. Development
120:2431–2442.
Polyhistidine Variants of HOXA1 in Neurons 487
Journal of Neuroscience Research DOI 10.1002/jnr