Direct Reprogramming of Huntington’s Disease Patient Fibroblasts into Neuron-Like Cells Leads to Abnormal Neurite Outgrowth, Increased Cell Death, and Aggregate Formation

Recent advances in trans-differentiation of one type cell to another have made it possible to directly convert Huntington’s disease (HD) patient fibroblasts into neurons by modulation of cell-lineage-specific transcription factors or RNA processing. However, this possibility has not been examined. Here, we demonstrate that HD patient-derived fibroblasts can be directly trans-differentiated into neuron-like cells by knockdown of the expression of a single gene encoding the polypyrimidine-tract-binding protein. The directly converted HD neuron-like cells were positive in expression of Tuj1, NeuN, DARPP-32, and γ-aminobutyric acid and exhibited neuritic breakdown, abnormal neuritic branching, increased cell death, and aggregation of mutant huntingtin. These observations indicate that the neuron-like cells directly converted from HD patient fibroblasts recapitulate the major aspects of neuropathological characteristics of HD and thus provide an additional model for understanding the disorder and validation of therapeutic reagents.


Introduction
Huntington's disease (HD) is a progressive neurodegenerative disorder caused by expansion of polyglutamine (polyQ) repeats in the N-terminus of the huntingtin (Htt) protein [1,2]. The disease is neuropathologically characterized by neuronal loss in the striatum and cortex and formation of protein aggregates (inclusions), resulting in motor and behavioral dysfunction [3]. To understand the pathogenesis of HD, a number of HD cell models have been created and applied in many studies over the last two decades [4,5]. Although these HD cells exhibit at least some of the pathological features of HD, most of them do not express fulllength human mutant Htt and neuronal markers and thus are not ideal for modeling HD. Induced pluripotent stem cells from HD patient or animal fibroblasts provide a new model for studying HD [6][7][8][9]. However, the neuronal induction process is usually timeconsuming and tedious. Recently, trans-differentiation of one type cell to another has been made it possible to directly convert HD patient fibroblasts into neuron-like cells by modulation of celllineage-specific transcription factors or RNA processing [10][11][12]. However, it remains unknown whether HD patient-derived fibroblasts can be directly reprogrammed into the neuron-like cells that reproduce the major aspect of HD pathological features.
The polypyrimidine-tract-binding (PTB) is an RNA-binding protein that regulates RNA splicing, stability, and localization [13]. During neuronal differentiation, the expression of PTB is switched to its neuronal homolog, nPTB [14]. Forced expression of PTB blocks neuronal differentiation [15], whereas knockdown of PTB expression by PTB-RNA interactions dramatically promotes conversion of diverse cell types into neurons [12,16]. Here, we demonstrate that following PTB knockdown, HD patient-derived fibroblasts can be directly reprogrammed to neuron-like cells that exhibit the major HD pathological characteristics.

Ethics statement
The following cell lines were obtained from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research: AG07095, GM04281, and GM05539. The Coriell Institute and ATCC maintain the written consent forms and privacy of the donors of the fibroblast samples, and the authors had no contact or interaction with the donors. All human fibroblast cells and protocols in the present study were carried out in accordance with the guidelines approved by the University of South Dakota Institutional Review Board.
Cell culture, preparation and infection of PTB1 smallhairpin (sh) RNA lentiviral particles Human fibroblasts were maintained in DMEM supplemented with 10% defined FBS, non-essential amino acids, Glutamax, bmercaptoethanol and 100 ng/mL bFGF at 37uC, 5% CO 2 . The CAG repeat number information in the htt gene was obtained from Coriell and confirmed by PCR using a PCR kit (Genelink).
Preparation of lentiviral particles of the shRNAs against human PTB1 and infection of fibroblasts were performed as previously described [12]. Sixteen hours after the shRNA treatment, the cells were selected either with 2 mg/ml puromycin or 100 ng/ml of hygromycin B for 48 h. Selected cells were switched into N3 medium (DMEM/F12, 25 mg/ml insulin, 50 mg/ml human transferrin, 30 nM sodium selenite, 20 nM progesterone, and 100 nM putrescine) supplemented with FGF2 (10 ng/ml) for 3 days and then switched to N3 medium for 10 days. Finally, cells were maintained in N3 medium supplemented with BDNF, GDNF, NT3 and CNTF as previously described [12] until being used for different analyses.

Cell counting
If a Tuj1-positive cell had lost all neurites or showed neurite breakdown, the cell would be treated as a cell with neuritic degeneration. Tuj1-positive cells with less than 20 mm in length of neurites or showing apparent thin neurites were regarded as cells with abnormal neuritic branching. GABA-positive cells with neurite breakdown and/or shrunken nuclei/cell bodies were counted as degenerated cells. If a cell had a nucleus containing one or more Htt aggregates, the cell would be counted as the positive for nuclear inclusion. If a cell contains aggregate(s) in the nonnuclear soma region or inside a neurite, the cell would be counted as the positive for non-nuclear (soma/neuropil) aggregate. At least 50 cells were counted in each experiment group and three independent experiments were performed.

Detection of apoptotic cells
Apoptotic cell death was examined as previously described [20] by utilizing a TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling, TUNEL) based apoptosis detection kit (Millipore). Stained cells were observed with a fluorescence microscope. Apoptotic cell rate was calculated as follows: apoptotic (or TUNEL positively stained cell) rate (%) = number of TUNELpositively stained cells/number of total cells (assessed by Hoechst 33342 staining)6100%.

Statistical analysis
Statistical comparisons between two groups were evaluated using two-tailed student's t test. P,0.05 was regarded as statistically significant.

Results and Discussion
To reprogram the fibroblasts derived from HD patients into neuron-like cells, we employed a recently described method to knock down PTB protein [12] by infecting HD patient fibroblasts expressing Htt containing either 16Q, 68Q, or 86Q with lentiviral shRNAs against human PTB. Nineteen days following PTB knockdown, the cells exhibited a typical neuron-like morphology and showed positive immunoreactivity with Tuj1, a neuronspecific cytoskeleton protein present in newly generated immature postmitotic neurons and differentiated neurons [20,21] (Fig. 1A). Cell counting showed that the HD patient-derived fibroblasts did not significantly differ from the normal fibroblasts in the capability of conversion to Tuj1-positive neuron-like cells (Fig. 1B). Those undifferentiated cells did not show Tuj1 staining and only showed the nuclear staining (Fig. 1A). The Tuj1-positive cells converted from HD patients (referred to as 68Q and 86Q, respectively) showed different neuritic morphology from the cells derived from a normal individual (16Q). The normal fibroblast-converted cells extended from one to several relatively thick neurites directly from the cell body (Fig. 1A, left panel). In addition to the thick neurites, however, the HD fibroblasts-derived cells frequently grew out thin neurites either directly from their cell bodies or from thick neurites (Figs. 1A, middle and right panels, pointed by arrow heads). Interestingly, some of the neurites derived from HD neuron-like cells were broken down and degenerated into small fragments positive in Tuj1 staining (Fig. 1A, pointed by arrows). Additionally, cell counting results indicated that more HD neuron-like cells (68Q and 86Q) exhibited abnormal neuritic branching (Fig. 1C) and neuritic breakdown (Fig. 1D) than the wild-type of neuronlike cells (16Q). These results indicate that the trans-differentiation of HD patient's fibroblasts into neuron-like cells leads to abnormal neuritic branching and degeneration.
As the neuronal nuclear antigen (NeuN) is a nuclear protein widely expressed in the mature postmitotic neurons, it has been commonly used as a neuron-specific marker for mature neurons [22]. We thus stained the cells with a NeuN specific antibody and found that at least 10% cells were positive in NeuN expression in each of the three converted cell types after nineteen days of the reprogramming (Fig. 2A). Since one major pathological feature of HD is selective loss of GABAergic neurons in the striatum [23], we next examined whether the trans-differentiated neuron-like cells express c-aminobutyric acid (GABA), an inhibitory neurotransmitter. As shown in Fig. 2B, GABA was strongly expressed in both the normal and HD neuron-like cells nineteen days after shRNA knockdown of PTB. Compared to the normal fibroblast-derived GABA-positive cells, some HD GABA-positive cells showed degenerating neurites and shrunken cell bodies (Fig. 2B, right  panel). Cell counting indicates that neurodegeneration was significantly more in the HD neuron-like cells than in the normal cells (Fig. 2C). Additionally, as degenerated neurons in HD striatum are DARPP-32 positive cells [23], we examined whether the trans-differentiated neuron-like cells are also positively stained with the protein. As shown in Figs. 2D and 2E, thirty days following the reprogramming, many cells expressed DARPP-32. At this time point, however, degenerated cells were dramatically increased to 59% and 79% in the 68Q and 86Q HD cells, respectively (Figs. 2F, 2G). Taken together, these data suggest that the HD patient fibroblasts can be trans-differentiated to GABA and DARPP-32-positive neuron-like cells and the reprogramming triggers increased cell death in the HD fibroblast-derived cells.
We next examined whether trans-differentiation of HD patient fibroblasts to neuron-like cells leads to mutant Htt aggregation. We therefore immunostained the three types of converted neuronlike cells with the well-documented EM48 Htt antibody, which selectively binds to the toxic N-terminal fragment of the mutant Htt protein [24], and then assessed the cells positive with inclusions in the nucleus, soma, and neuropil by confocal microscopy. There was no EM48-positive nuclear inclusion in the neuron-like cells trans-converted from the normal fibroblasts, whereas most of the neuron-like cells converted from the two types of HD fibroblasts contained Htt inclusions in their nuclei and nonnuclear regions (soma and neuropils) (Figs. 3A, 3B). These results indicate that the mutant Htt proteins preferentially form aggregates in both the nucleus and non-nuclear regions upon conversion to neuron-like cells.
One interesting observation from this research is that direct conversion of HD patient fibroblasts to neuron-like cells leads to abnormal neurite outgrowth and branching, characterized by frequently short or thin neurite outgrowth. This is in accordance with a previous in vivo study, in which abnormal dendritic arbors and increased dendritic branching in spiny striatal neurons were identified in post-mortem HD patients' brain sections [25]. Although it remains unclear why the HD neuron-like cells selectively exhibit this dysmorphic alteration, mutant Htt-caused intracellular trafficking dysfunction may be, at least partially, responsible for abnormal neurite outgrowth and branching [26]. Additionally, mutant Htt also impairs mitochondrial integrity [27] and disrupts production and trafficking of neurotrophic factors [28], which may also affect neurite outgrowth and branching. As a further direction, it is interesting to explore the biological significance underlying this dysmorphic alteration. In addition to showing increased cell death, the directly trans-converted HD cells also form aggregates not only in the nucleus but also in nonnuclear regions such as neuropils, which is in accordance with previous in vivo studies using HD patient brain tissues [29]. Thus, the directly converted neuron-like cells from HD fibroblasts provide a reliable model for studying pathogenic mechanisms of HD and may be a useful tool for validation of therapeutic target or drugs in the future.