ABSTRACT
Despite the high incidence of severe defects in the central nervous system caused by human cytomegalovirus (HCMV) congenital infection, the mechanism of HCMV neuropathogenesis and the roles of individual viral genes have not yet been fully determined. In this study, we show that the immediate-early 2 (IE2) protein may play a key role in HCMV-caused neurodevelopmental disorders. IE2-transduced neural progenitor cells gave rise to neurospheres with a lower frequency and produced smaller neurospheres than control cells in vitro, indicating reduction of self-renewal and expansion of neural progenitors by IE2. At 2 days after in utero electroporation into the ventricle of the developing brain, a dramatically lower percentage of IE2-expressing cells was detected in the ventricular zone (VZ) and cortical plate (CP) compared to control cells, suggesting that IE2 concurrently dysregulates neural stem cell maintenance in the VZ and neuronal migration to the CP. In addition, most IE2+ cells in the lower intermediate zone either showed multipolar morphology with short neurites or possessed nonradially oriented processes, whereas control cells had long, radially oriented monopolar or bipolar neurites. IE2+ callosal axons also failed to cross the midline to form the corpus callosum. Furthermore, we provide molecular evidence that the cell cycle arrest and DNA binding activities of IE2 appear to be responsible for the increased neural stem cell exit from the VZ and cortical migrational defects, respectively. Collectively, our results demonstrate that IE2 disrupts the orderly process of brain development in a stepwise manner to further our understanding of neurodevelopmental HCMV pathogenesis.
IMPORTANCE HCMV brain pathogenesis has been studied in limited experimental settings, such as in vitro HCMV infection of neural progenitor cells or in vivo murine CMV infection of the mouse brain. Here, we show that IE2 is a pivotal factor that contributes to HCMV-induced abnormalities in the context of the embryonic brain using an in utero gene transfer tool. Surprisingly, IE2, but not HCMV IE1 or murine CMV ie3, interferes pleiotropically with key neurodevelopmental processes, including neural stem cell regulation, proper positioning of migrating neurons, and the callosal axon projections important for communication between the hemispheres. Our data suggest that the wide spectrum of clinical outcomes, ranging from mental retardation to microcephaly, caused by congenital HCMV infection can be sufficiently explained in terms of IE2 action alone.
INTRODUCTION
Human cytomegalovirus (HCMV) is a member of the betaherpesviruses that infect most human populations in the world and establishes a lifelong infection, with cycles of repeating latency and reactivation (1). Although HCMV infection is mostly asymptomatic or subclinical in immunocompetent individuals, it can cause morbidity and mortality in immunocompromised patients (1). Because of widespread infection in human populations, the absence of an effective vaccine, and easy mother-to-fetus viral transmission, HCMV is the most common cause of congenital infections (2) and also the leading viral cause of birth defects, affecting 0.25% to 2% of newborn infants worldwide (3).
Up to 10% of congenitally HCMV-infected newborns have overt clinical manifestations at birth, including microcephaly, thrombocytopenia, jaundice, hepatomegaly, and occasionally life-threatening systemic complications (4). The remaining infected infants have no clinical manifestations at birth, but some will still have defects in brain functions, such as mental retardation, visual and hearing disorders, seizures, and epilepsy (3, 5, 6). Infections during early pregnancy generally lead to more severe clinical manifestations in newborns (5).
For proper brain development, neural stem cell number and expansion must be closely regulated. Embryonic neural stem cells are maintained in the ventricular zone (VZ), and as brain development proceeds, they differentiate into cells with neuronal lineage commitment and migrate away from the VZ to the outer surface of the brain to form cortical lamination (7). The failure of precise regulation of neural stem cell proliferation and differentiation causes developmental abnormalities in the brain such as microcephaly and mental retardation (8).
HCMV brain pathogenesis has been studied mainly in cultured neural progenitor cells in vitro because of the lack of an appropriate animal model due to the strict species-specific tropism of HCMV that only allows virus replication in cells of human origin (9). To circumvent this limitation, an animal model using murine CMV (MCMV) infection of the mouse brain has been used to provide insight into CMV pathogenesis in the central nervous system (CNS) (10, 11). However, since MCMV differs considerably from HCMV with respect to immediate-early (IE) gene organization, overall nucleotide sequence, and corresponding biological properties (12), there have been limitations in extrapolating data from the experimental animal MCMV model to HCMV pathogenesis in the human brain.
To better understand the pathogenesis of viruses with a large and complicated genome like HCMV, in-depth studies analyzing cellular and molecular changes at a single-gene level as well as the whole-virus level are required. Considering that gene function studies inducing expression of a viral gene of interest by vectors circumvent the issue of host cell tropism, it is surprising that few attempts, to our knowledge, have been made to assess the effect of individual HCMV genes in vivo using an animal model.
In this study, we show that expression of the HCMV immediate-early 2 (IE2) protein has a profound negative impact on the proliferation and self-renewal capacity of neural progenitor cells in vitro. Expression of HCMV IE2 protein also induces premature differentiation of neural stem cells and accumulation of nonfunctional immature neurons during brain development in vivo. Our study provides insight into the molecular pathogenesis mechanisms by which a complex virus, HCMV, disturbs the development of the most complex organ, the brain.
RESULTS
Expression of HCMV IE2, but not IE1, impairs the self-renewal and proliferation of neural progenitor cells in vitro.For efficient gene delivery into primary neural progenitor cells, IE1 and IE2 genes were cloned into the retroviral vector MSIG (13) (Fig. 1A). MSIG contains an internal ribosome entry site (IRES) sequence for bicistronic expression of IE1 or IE2 and green fluorescent protein (GFP) as a reporter. Western blotting (Fig. 1B) and immunofluorescence (Fig. 1C) showed that the constructs expressed proteins of the expected sizes for each IE protein and GFP in target cells.
HCMV IE2 expression, but not IE1 expression, negatively regulates neural stem cells in vitro. (A) Schematic representation of the retroviral vector used in this study. The retroviral vector MSIG, which bicistronically expresses HCMV IE and GFP, was used to transduce primary embryonic neural progenitor cells. MSIG expressing only GFP was used as a control. IRES, internal ribosome entry site; LTR, long terminal repeat; MCS, multicloning site; MIE, major immediate early. Expression of IE1/2 and GFP in transduced target cells was confirmed by Western blotting using an anti-IE antibody (B) and fluorescence microscopy (C), respectively. (D) Effect of IE1 and IE2 on neurosphere formation. E14.5 primary neural progenitors transduced with each retroviral vector were cultured for 7 days in serum-free medium supplemented with B27/N2 and FGF2. (E) Quantification of neurosphere frequency and size from panel D. (F) Test for possible cytotoxic effects of IE proteins. Cell lysates from primary neural progenitors transduced with each retroviral vector and cultured for 2 days were tested for cleaved caspase-3 generation by Western blotting. Staurosporine (STS; 2 μM; 8 h of incubation) was used as a positive control. Scale bars, 100 μm. Error bars represent standard deviations (SD). Student's t test was used to determine statistical significance. ***, P < 0.001; n ≥ 3.
To test the effect of IE proteins on embryonic neural progenitor cells in vitro, we began with neurosphere cultures using mouse embryonic day 14.5 (E14.5) neural progenitors. The neurosphere assay is a clonal cell survival assay based on the ability of a single neural stem cell to grow into a colony (neurosphere) in defined stem cell media (14). By measuring neurosphere frequency and size, the effects of a gene of interest on self-renewal and proliferative activities, respectively, can be determined for stem cells. As shown in Fig. 1D and E, IE2-transduced cells gave rise to far lower frequencies of neurospheres and smaller neurospheres than control cells, indicating reduction of self-renewal and cell proliferative capabilities upon IE2 expression, whereas IE1 expression did not appear to have any deleterious effects on neurosphere formation. To rule out the possibility that IE2 exerts its negative influences on neural progenitor cells through cytotoxic activity, we performed Western blotting to detect the active form of caspase-3, a marker for apoptosis, and found that cleaved caspase-3 was undetectable in IE2-transduced neural progenitor cells (Fig. 1F). Taken together, these results indicate that IE2 negatively regulates neural stem cell characteristics in vitro, and these observations were not due to increased cell death.
IE2 negatively acts on multiple steps of brain development in vivo.We next tested the effects of IE2 expression on neural stem cells in vivo. Neural stem cells reside and proliferate in the VZ throughout brain development, and cells undergoing differentiation into neurons migrate away from the VZ to cross the intermediate zone (IZ) and reach the cortical plate (CP). For in vivo transgene expression in the developing brain, IE2-expressing plasmid DNA was introduced into the telencephalic ventricle of E13.5 embryos by in utero electroporation (Fig. 2A), and the outcomes of IE2 expression were analyzed at E15.5. Initially, only cells lining the ventricles (thus, mostly neural stem cells) are transfected by electroporation, and after a certain period of time, the distribution of gene-transferred cells in the cortical laminar layers are visualized by immunofluorescence against the GFP reporter protein to analyze the differentiation pattern of transgene-expressing cells. IE and GFP expression levels after in utero electroporation were verified by immunofluorescence using anti-IE1/2 and anti-GFP antibodies (Fig. 2B). As shown in Fig. 2C and D, IE2 expression resulted in significant changes in cell localization in the embryonic brain. At 2 days postelectroporation, a dramatically higher fraction of IE2-transfected cells was detected in the multipolar morphology zone (MMZ) than in control cells. Far lower percentages of IE2-expressing cells in the VZ and CP indicate that IE2 dysregulates neural stem cell maintenance as well as neuronal migration and maturation. Enhanced cell exit from the VZ upon IE2 expression was further confirmed by immunostaining using an anti-Sox2 antibody (Fig. 2E). The delay in migration of IE2-expressing cells toward the pial surface was still obvious at E18.5 and thus was presumed to be maintained throughout brain development (Fig. 2H). The presence of a large number of IE2-expressing cells in the transfected brain up to 5 days postelectroporation also provides evidence for the cytotoxicity-independent dysregulation of brain development by IE2.
HCMV IE2 expression disturbs cortical brain development in vivo. (A) Schematic of gene delivery into the embryonic brain through in utero electroporation of plasmid DNA and assessment of the neurodevelopment-modulating activity of a gene of interest by determining the transgene-expressing cell positions in the cortical layers. Initially, only neural stem cells lining the VZ are electroporated, and as development proceeds, they give rise to neural stem cells that reside in the VZ or neurons that migrate up to the pial surface. (B) IE expression in the brains at 2 days postelectroporation was confirmed by immunostaining using anti-GFP (reporter gene; green) and anti-IE1/2 (red) primary antibodies as well as Alexa Fluor 488- and 555-conjugated secondary antibodies. IE2-expressing cells in E15.5 (C) or E18.5 (H and I) embryonic brains that were electroporated at E13.5 were immunolabeled using anti-GFP primary and Alexa 488-conjugated secondary antibodies. For panel C, GFP immunofluorescence merged with DAPI-counterstained images are shown on the right. (D) Quantification of GFP+ cell positions from panel C. (E) Double immunolabeling of E15.5 brain sections electroporated with IE2-expressing plasmid at E13.5 using anti-GFP (green) and anti-Sox2 (red) primary antibodies. (F) The shapes of GFP+ electroporated cells in E15.5 cortices at the transition between MMZ and RMZ. The dotted red lines indicate the direction of radial migration. Radial cells were defined as cells with a leading process with a deviation angle of less than 45° relative to the normal radial migration direction. (G) Quantification of cell shapes shown in panel F. (I) Examination of callosal axon trajectory (closed arrowheads) using immunolabeling of GFP at E18.5. The dotted lines indicate the midline, and an open arrowhead indicates where IE2+ callosal axons stop projection. VZ, ventricular zone; SVZ, subventricular zone; IZ, intermediate zone; CP, cortical plate; MMZ, multipolar morphology zone; RMZ, radial morphology zone; CC, corpus callosum; LV, lateral ventricle. Scale bars were 20 μm (B, E, and F), 50 μm (C), and 100 μm (H and I). Error bars represent SD. Student's t test was used to determine statistical significance. **, P < 0.01; ***, P < 0.001; n ≥ 3.
In addition, there was a prominent difference in cellular morphology between IE2+ and control cells in the boundary region between the MMZ and radial morphology zone (RMZ), where multipolar-to-bipolar transition occurs. Control cells had long, radially oriented monopolar or bipolar neurons, whereas most IE2+ cells either showed multipolar morphology with short neurites or possessed a leading process deviating from the normal radial migrating direction (Fig. 2F). To quantify these observations, the deviation angle of the leading process relative to the normal radial migration direction was measured, and cells with a deviation angle of less than 45° were defined as radial cells and the others as nonradial cells (Fig. 2G).
The neocortical hemispheres are connected via a large axon bundle, the corpus callosum. This connection is achieved by the orderly projection of axons over long distances from one side of the cortex to the equivalent contralateral cortical area (15). As shown in Fig. 2I, callosal axons of IE2-expressing cells failed to project into the topographically equivalent locations in the contralateral brain area at E18.5. Taken together, these results indicate that IE2 disrupts multiple processes of brain development, including neural stem cell maintenance in the VZ, migrating neuronal polarization, radial migration, and callosal axon projection linking the two hemispheres.
Expression of MCMV ie3, the MCMV homolog of HCMV IE2, has no detrimental effects on neural stem regulation.Because of the high level of conservation between HCMV IE2 and MCMV ie3 with regard to amino acid sequence, gene structure, and protein functions (16), we tested if expression of MCMV ie3 could exert negative effects similar to those of HCMV IE2 in neural stem cell regulation. Interestingly, MCMV ie3 behaved like the negative control both in vitro and in vivo; MCMV ie3 expression appeared to have no inhibitory effects on the formation rate and size of neurospheres (Fig. 3B and C). The cell distribution pattern in the developing brain also was not affected by ie3 expression induced by in utero electroporation (Fig. 3E and F). Taken together, these results indicate that despite its similarity to HCMV IE2, MCMV ie3 is not associated with neural stem cell misregulation.
Effects of MCMV ie3 expression on neural stem cell regulation. (A) Expression of Myc-tagged MCMV ie3 in retrovirally transduced embryonic neural progenitors was confirmed by Western blotting using an anti-Myc tag antibody. (B) E14.5 primary neural progenitors transduced with a retroviral vector expressing either HCMV IE2 or MCMV ie3 were cultured for 7 days in neural stem cell medium for neurosphere formation. (C) Quantification of neurosphere frequency and size in of cells shown in panel B. (D) MCMV ie3 expression in the brains at 2 days postelectroporation was examined using anti-GFP (green) and anti-Myc tag (red) primary antibodies and Alexa Fluor 488- and 555-conjugated secondary antibodies. (E) MCMV ie3-expressing cells in E15.5 embryonic brains that were electroporated at E13.5 were immunostained using anti-GFP primary and Alexa Fluor 488-conjugated secondary antibodies. DAPI-merged images are shown on the right. (F) Quantification of GFP+ cell positions of cells shown in panel E. Scale bars were 25 μm (D), 50 μm (E), and 100 μm (B). **, P < 0.01; ***, P < 0.001; n ≥ 3.
IE2 expression dysregulates neural progenitor cells in vitro through cell cycle-inhibitory and DNA binding activities.To address how IE2 perturbs neural stem cell properties, we generated two functional mutant IE2 genes by site-directed mutagenesis. IE2 induces transactivation of target genes and host cell cycle arrest, and these functions can be selectively abrogated by amino acid substitutions. The H446,452L mutation in a putative zinc finger domain abolishes both DNA binding and host cell cycle arrest abilities (17–19), and the Q548R mutation, which adds a positively charged residue in close proximity to the acidic activation domain, disrupts only the latter (20) (Fig. 4A). The mutant IE2 genes were expressed at levels comparable to those of wild-type IE2 in target cells as assessed by Western blotting (Fig. 4B). Changes in host cell cycle-inhibiting and DNA binding activities of IE2 by the mutations were then tested by quantitative real-time PCR (qPCR) to detect expression of cell cycle inhibitor genes, cdkn1a (p21) (21) and cdkn1b (p27) (22), in the mouse neural progenitor cells (Fig. 4D) and a chromatin immunoprecipitation (ChIP) assay to test IE2 binding to the c-myc promoter region using chromatin extracts derived from mouse neuroblastoma Neuro-2a cells (Fig. 4E).
Effects of IE2 mutations on IE2-induced downregulation of neural stem cell properties in vitro. (A) Schematic diagram showing the location of the H446,452L and Q548R mutations in the IE2 gene. AD, activation domain; CORE, core domain; N, nuclear localization signal; ZF, putative zinc finger motif. Expression of IE2 mutants and GFP in transduced target cells was confirmed by Western blotting using an anti-IE antibody (B) and fluorescence microscopy (C). (D) Two days after E14.5 primary neural progenitor cells were transduced with each retroviral vector, mRNA expression levels of the indicated cell cycle inhibitor genes were measured by qPCR. (E) Chromatin immunoprecipitation (ChIP) assay using Neuro-2a (a murine neuroblastoma cell line) chromatin extracts and anti-Myc tag antibody and subsequent qPCR analysis for the genomic c-fos promoter region. (F) Effect of IE2 mutants on neurosphere frequency and size. E14.5 primary neural progenitors transduced with the retroviral vectors were cultured for 7 days in serum-free medium supplemented with B27/N2 and FGF2. (G) Quantification of results depicted in panel F. Scale bars, 100 μm. Error bars represent SD. Student's t test was used to determine statistical significance. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n ≥ 3.
Upon transduction into primary neural progenitor cells, the Q548R mutation restored neurosphere frequency as well as neurosphere size to 50% of the wild-type IE2 level (Fig. 3F and G), suggesting that the reduced level of in vitro stem cell properties induced by IE2 expression is partly attributable to cell cycle inhibition. The DNA binding activity of IE2 appeared to account for the remaining portion of the inhibitory effects of IE2 on neurosphere formation, since the H446,452L mutation almost completely abolished the ability of IE2 to reduce neurosphere frequency and size. These results indicate that the IE2 cell cycle inhibition and DNA binding activities comparably contribute to IE2-induced reduction of neural progenitor cell stemness in vitro.
IE2 cell cycle inhibition and DNA binding abilities are responsible for accelerated cell exit from the VZ and mingled neuronal cell polarity in vivo, respectively.Finally, we examined the functional consequences of the IE2 mutations in vivo by in utero