Human cytomegalovirus induces significant structural and functional changes in terminally differentiated human cortical neurons

ABSTRACT Human cytomegalovirus (HCMV) is a highly prevalent viral pathogen that typically presents asymptomatically in healthy individuals despite lifelong latency. However, in 10%–15% of congenital cases, this beta-herpesvirus demonstrates direct effects on the central nervous system, including microcephaly, cognitive and learning delays, and hearing deficits. HCMV has been widely shown to infect neural progenitor cells, but the permissiveness of fully differentiated neurons to HCMV is controversial and chronically understudied, despite potential associations between HCMV infection and neurodegenerative conditions. Using a model system representative of the human forebrain, we demonstrate that induced pluripotent stem cell-derived excitatory glutamatergic and inhibitory GABAergic neurons are fully permissive to HCMV with complete viral replication, competent virion production, and spread within the culture. Interestingly, while cell proliferation was not induced in these post-mitotic neurons, HCMV did increase expression of proliferative markers Ki67 and proliferative cell nuclear antigen suggesting alterations in cell cycle machinery. These findings are consistent with previous HCMV-mediated changes in various cell types and implicate the ability of viral proteins to alter proliferative pathways to promote virion production. Infection also induces significant structural changes in forebrain neurons, such as the formation of syncytia and the retraction of neurites. Finally, we demonstrate that HCMV disrupts calcium signaling and decreases neurotransmission, with action potential generation effectively silenced after 15 days post-infection. Taken together, our data highlight the potential for forebrain neurons to be permissive to HCMV infection in the central nervous system, which has significant implications for overall brain health and function. IMPORTANCE Human cytomegalovirus (HCMV) is a highly prevalent viral pathogen that can cause serious neurological deficits in infants experiencing an in utero infection. Also, as a life-long infection, HCMV has been associated with several diseases in the adult brain. HCMV is known to infect early neural progenitor cells, but whether it also infects terminally differentiated neurons is still debated. Here, we differentiated human-induced pluripotent stem cells into neurons for 84–120 days to test the ability of HCMV to infect terminally differentiated neurons and assess the downstream functional consequences. We discovered that mature human neurons are highly permissive to HCMV infection, exhibited late replication hallmarks, and produced infectious virus. Moreover, infection in terminally differentiated neurons essentially eliminated neuron function. These results demonstrate that terminally differentiated human neurons are permissive to HCMV infection, which can significantly alter both structural and functional features of this mature neuron population.

KEYWORDS human cytomegalovirus, neurons, forebrain, induced pluripotent stem cells, iPSCs, function, structure, calcium signaling, Ki67, cell cycle, synapse, multielectrode array H uman cytomegalovirus (HCMV) is a common pathogen with a worldwide infection rate ranging from 60% to 90% of the population (1)(2)(3).Though immunocompetent individuals are typically asymptomatic, seropositive immunocompromised individuals have a high rate of significant disease.Upon vertical transmission, a subset of congenital infections (10%-15%) demonstrate a risk to fetal health and development (4)(5)(6)(7)(8)(9).In such cases, a wide range of neurological symptoms may occur, including hearing impairment, cognitive deficits, learning/language disorders, and microcephaly (2,(9)(10)(11).HCMV's prominence is such that it currently is the most common infectious cause of birth defects in the United States (12).Interestingly, several studies have also examined neurological effects present in individuals with asymptomatic congenital infections (2,4,13,14).Symptoms in immunocompromised individuals are significantly more pronounced and can require hospitalization due to altered liver function, neurological conditions, and immune responses (15).Together, these studies highlight the negative impact of HCMV infection on the development and activity within the central nervous system (CNS).
The role of HCMV within the CNS is complex and multifaceted.Previous reports demonstrate that HCMV can infect several resident cell types within the CNS, includ ing astrocytes (16)(17)(18), oligodendrocytes (19), and neural progenitor cells (NPCs) (7,(20)(21)(22)(23).However, the ability of neurons to act as a susceptible target is often deba ted, with empirical results ranging from no detected infectivity to full permissiveness (16,(22)(23)(24)(25)(26)(27).Among those demonstrating a lack of HCMV replication within neurons, the cells' post-mitotic state is commonly indicated to be the underlying reason (16).Functionally, little is known about the effects of HCMV on terminally differentiated neuronal populations, though early reports detail tendencies toward cell death and altered responses to glutamate stimulation (27).This effect is in addition to functional aberrancies observed in other cell types such as NPCs and fibroblasts, often involving dysregulation of calcium signaling (28)(29)(30).Considering the limitations of existing data, further analysis of the effect of direct infection on neurons is warranted.
Human induced pluripotent stem cells (iPSCs) represent an indispensable tool for studying the effects on human tissues that are otherwise not accessible.This is evident through both their robust differentiation capabilities-allowing for the study of the viral effects on various cell types-and their species compatibility with HCMV.To date, a majority of studies analyzing the effect of HCMV infection on CNS cells have been conducted using either fetal-derived primary cultures or cell type-specific lines (7,16,17,19,21,22,31,32), while, in contrast, few have utilized iPSC-derived cultures (27,28,33).Furthermore, iPSCs allow for convenient, accurate modeling of neuronal populations present in the human forebrain, an area that is both responsible for higher cognitive functions and influenced by congenital HCMV-mediated neurological symptoms.The iPSC-derived neurons may be particularly relevant for studying the role of HCMV in neurodegenerative conditions such as Alzheimer's disease (AD) (34)(35)(36).
In the current study, the structural and functional effects of HCMV infection were studied in the context of iPSC-derived, terminally differentiated forebrain neurons.We assessed HCMV replication from immediate-early gene expression through virion production in terminally differentiated and functionally mature iPSC-derived excitatory and inhibitory neurons.We discovered that these mature neurons are highly permissive to HCMV infection, exhibited late replication hallmarks such as expression of the late protein pp28, and produced infectious virus.Finally, we show that HCMV infection in terminally differentiated neurons essentially eliminated action potential generation and calcium signaling.These results demonstrate that terminally differentiated human neurons are permissive to HCMV and that infection significantly alters both structural and functional features of mature neurons.As such, these data suggest that HCMV infection could have pathological consequences in the adult CNS.

Terminally differentiated forebrain neurons derived from iPSCs support HCMV lytic replication
Utilizing iPSCs from five unrelated adult individuals (three healthy control lines, an AD patient line with a PSEN2 N141I mutation, and an AD patient with sporadic AD [sAD]), forebrain neurons were differentiated through a neural progenitor cell (NPC) intermedi ate (Fig. 1A).As HCMV presence has been documented to inhibit neural differentiation (7,21), cultures were propagated for 84-150 days of differentiation (assay dependent) prior to infection to ensure the presence of functionally mature neurons in culture.Neuronal cultures derived from fetal cortex progenitors were used in tandem to validate findings and limit potential concerns in using this iPSC approach.Fetal cells were cultured for ~20 weeks as non-adherent neurospheres (37,38) and then plated as a monolayer and differentiated for an additional 30 days (Fig. 1B).iPSC-derived forebrain neurons examined at day 84 of differentiation show the dominant presence of both inhibitory GABA-ergic (GABA + ) neurons and excitatory glutamatergic (VGLUT2 + ) neurons (Fig. 2A).Additionally, both vimentin + astrocytes and minute quantities of Ki67 + proliferative cells (likely NPCs) can be found within the iPSC-derived cultures (Fig. 2A).In iPSC-derived cultures, neurons comprised 62.61% of the total cells in culture with glial cells con stituting the remaining population (Fig. 2B).Fetal progenitor-derived cultures also contained high numbers of GABA-ergic and glutamatergic neurons (Fig. 2C), highlight ing consistency between models.Cell distributions in fetal-derived cultures closely mirrored those in their iPSC-derived counterparts, further demonstrating similarity between model systems (Fig. 2D).Mirroring in vivo neuronal maturation, we observe a change in cell morphology over time within iPSC-derived cultures.NPCs exhibit a more cobblestone-like appearance and lack defined neurite projections, while mature neurons possess greater neurite elongation, development, and network establishment in late cultures (Fig. 3).As expected, we observe a reduction in expression of the NPC marker Pax6 and a near complete loss of Ki67 expression over the differentiation timeline (Fig. 3).These facets of neuronal development are mirrored in the fetal-derived cells (Fig. 3).Together, these data establish the iPSC model system and clearly demonstrate the presence of post-mitotic differentiated neurons.
With the culture system established, we next infected D30 forebrain neurons with HCMV originating from the BAC-cloned TB40/E strain.We used two recombinant versions, HCMV TB40/E-eGFP, containing the eGFP gene under control of an SV40 promoter (39,40), and TB40/E-eGFP/mCh expressing late protein pp28 in-frame with the fluorescent protein mCherry and IE2 in-frame with a cleavable eGFP, IE2-2A-eGFP UL99-mCh (TB40/E-eGFP/mCh) (41) (Fig. 4A).The latter construct was chosen to demonstrate progressive infection (TB40/E-eGFP/mCh) in select scenarios, while the former construct was more adaptable regarding fluorescence-based assays, such as microscopy (specifically in multichannel imaging) and immunoblotting.No phenotypic differences were noted between the two constructs.To test susceptibility of mature neurons, we infected at 0.5, 1, and 3 IU/cell (Fig. 4B and C).Virus was added to forebrain neurons for 2 hours, after which the inoculum was removed, cells washed with PBS, and fresh medium added every other day for the duration of the experiment.Utilizing TB40/ E-eGFP/mCh, we observed fluorescence at all multiplicities of infection (MOIs) within 48 hours post-infection (hpi) with an MOI of 3 being the most robust (Fig. 4B).By 15 days post-infection (dpi), differences in fluorescence between MOIs appear negligible (Fig. 4C).Expression of pp28-mCherry in forebrain neuron cultures confirmed that infection was proceeding from the initial step of viral entry through late gene expression (Fig. 4B, C, and 5A).Due to culture heterogeneity, immunofluorescence was utilized to determine the cell type being infected.Using two neuron-specific structural proteins-ßIII-Tubulin (Tuj1) and Neurofilament 200 (NF200)-we demonstrate that markers of infection (eGFP) colocalize with canonical neuronal markers (Fig. 5B).Furthermore, to confirm infection in a second model system, fetal cortex-derived neurons were infected with TB40/E-eGFP at an MOI of 3 and demonstrated a progressive increase in the number of GFP + cells over time (Fig. 6).To evaluate viral spread in long-term forebrain neuron cultures, time course data were collected tracking infection from the point of infection (D77 of differentiation) through 7 dpi.This timepoint was originally selected as the ideal timepoint for investigation due to it being both long enough to view infection-related phenotypes in cultures and limiting potential cell death resulting from long-term infection.Robust eGFP and mCherry signals were detected over the duration of the experiment, with some cells demonstrating primarily late gene (pp28) expression by 169 hpi (Fig. 7A).In conjunction with imaging experiments, cell pellets from forebrain neuron cultures were collected at daily time points, processed via DNA extraction, and evaluated for relative viral-to-cellular DNA over time.Utilizing combined data from five iPSC lines, we observed a significant increase in the percentage of viral DNA between days 1 and 6 (Fig. 7B).Supporting these data, fluorescence tracking revealed the expression of IE2 (observable via eGFP) becoming prominent around 90 hpi.These results were notably consistent between the three controls and the PSEN2 N141I line, whereas the sAD line demonstrated a lower amount of eGFP fluorescence.Late gene expression (mCherry) was noted shortly thereafter, with fluorescence increasing near 100 hpi (Fig. 7C and D).All lines were consistent regarding the amount and rate of mCherry fluorescence over time.
After confirming that forebrain neurons could be infected by HCMV, we next asked if infected forebrain neurons could produce infectious virus.At D84, D114, and D144 of differentiation, cells were infected with TB40/E-eGFP for 2 hours, washed to clear inoculum, and cultured in fresh medium for 14 days.The referenced timepoints (deter mined by adding 30, 60 and 90 days to the minimum 24 days of differentiation required to produce forebrain neurons) were selected to determine how neuronal maturity affects viral production.Conditioned media from forebrain neurons were collected and applied to ARPE-19 epithelial cells to determine viral titers.Across all timepoints, cells from each of the iPSC lines produced infectious virus (Fig. 7E).Furthermore, no significant differences were found between cell lines at each timepoint.It should be noted that although viral titers were lower than those previously observed in fibroblasts (~10 5 to 10 6 infec tious units per milliliter [IU/mL]), neuronal titer data closely match those of infected epithelial cells (~10 3 to 10 4 IU/mL) (42,43).Notably, the production of infectious virus in neural cultures signifies viral permissiveness.HCMV titer quantities from infected iPSCderived neurons were also compared with those from both infected iPSC-derived NPCs and infected fetal cortex-derived neurons.iPSC-derived neurons demonstrated little difference as compared with iPSC-derived NPCs (Fig. 7E).Fetal-derived neurons were less consistent, with one of the two tested lines showing no viral titer (Fig. 7E).To further examine virus production, viral particle infectivity was assessed.We determined the level of released viral particles by collecting conditioned media, DNase treating, and quantify ing virion-protected viral DNA.We then calculated the ratio of released viral particles per released infectious virus.Again, regardless of timepoint and iPSC line, average particle infectivity ratios were relatively similar (Fig. 7F).Comparisons to NPCs and the permissive fetal-derived culture revealed no significant differences compared with forebrain neuron cultures (Fig. 7F); however, infectious particles generated by forebrain neurons in the culture media are lower than those observed following infection of MRC-5 fibroblasts (approximately 67 genomes/IU at 96 hpi) (41).Taken together, these data demonstrate that iPSC-derived forebrain neurons are susceptible to HCMV infection and support the full viral replication cycle.

Infection by HCMV stimulates markers of proliferating cells in post-mitotic forebrain neurons
Infection by HCMV manipulates cell cycle in proliferating human fibroblasts and epithelial cells (44)(45)(46)(47).Mature neurons are post-mitotic, having entered the G0 phase upon full differentiation from NPCs (48).As such, the effects of HCMV on the neuronal cell cycle are largely unknown.To determine if HCMV infection reverts differentiated neurons back into a mitotic state, we evaluated expression of several cell cycle proteins.Though not significant, western blot analysis of D98 forebrain neurons (14 dpi) highligh ted a potential trend toward HCMV-dependent decreases in mitotic Cyclin B in all lines, with three of the five tested lines approaching significance (Fig. 8A and B).We then analyzed the protein expression of cyclin-dependent kinase inhibitor p21, and infection did not appear to have any consistent effect on its expression across cell lines (Fig. 8C).Infection resulted in a trending increase in PCNA expression (Fig. 8D and E).PCNA is an essential protein in DNA synthesis and is often elevated in the S phase.We next assessed expression of a marker of proliferation, Ki67.Using immunofluorescence in D84 (7 dpi) cultures, significant increases in Ki67 staining were detected within three of the four tested forebrain neuron cultures (Fig. 8F and G).These results are consistent with previous studies using multiple cell lines and HCMV strains (49-51).Furthermore, Ki67 staining primarily colocalized with areas of infected cells (Fig. 8G, denoted by red arrows, Movie S1).Taken together, these data demonstrate that HCMV infection of post-mitotic forebrain neurons stimulates common markers of the S phase observed in proliferation cells.

Infection alters human forebrain neuron structure
We and others have noted an increase in cell death within human NPCs and cerebral organoids upon HCMV infection (21,52,53).Therefore, we next tested whether HCMVinduced cell death of terminally differentiated forebrain neurons.We used terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and found no significant increase in TUNEL-positive cells in the HCMV-infected cultures to mock controls at D84 (7 dpi) (Fig. 9A and B).These findings were replicated at 14 dpi (21.90% ± 3.05% mock vs 22.46% ± 6.09% HCMV), demonstrating a consistent lack of cell death within sustained cultures.
Neuronal infection with the alpha-herpesvirus, herpes simplex virus (HSV) has been shown to transport proteins and capsid particles from the cell body through neurites (47,(54)(55)(56).As such, we next examined pp28-mCherry localization in both the cell body and in neurite projections.We infected D84 forebrain neurons with TB40/E-eGFP/mCh and observed pp28-mCherry expression within neuronal projections and large accumula tions of pp28-mCherry within cell bodies at 9 and 15 dpi (Fig. 9C, affected projections denoted by blue arrows).pp28 is proposed to be trafficked to the cytoplasmic virion assembly compartment, a process necessary for virion assembly (57).We observed VAC formation in forebrain neuron cultures indicated by large, punctate regions of pp28 (Fig. 9D).
Syncytia formation occurs during HCMV infection whereby disparate cells fuse to become a single, multinucleated cell often with a central VAC (58)(59)(60).We observed that forebrain neurons infected with HCMV do fuse into syncytia-like formations as character ized by a pp28-positive VAC surrounded by a ring of nuclei (Fig. 9D).To confirm that pp28 localization was not an artifact of the pp28-mCherry fusion construct, we subsequently confirmed pp28 protein localization using immunofluorescence (Fig. 9E).These struc tures were also present in cultures infected with TB40/E-eGFP (Fig. 9F).To determine if the cells composing the syncytia were once individual neurons, we completed immuno fluorescence analysis using an antibody for Tuj1.Syncytia structures were positively stained for Tuj1 (Fig. 9F; Movie S1), although the morphology was no longer consistent with a terminally differentiated neuron.We observed similar structures in the fetal cortex-derived neurons (Fig. 9F).Taken together, these results demonstrate that infection by HCMV causes post-mitotic forebrain neurons to fuse and express markers of prolifera tion without signs of significant cell death.

HCMV infection damages forebrain neuron function
Previously, our group demonstrated that infection negatively impacts calcium signaling in fibroblasts, NPCs, and mixed neural populations from cortical organoids by depleting calcium stores and compromising ATP/KCl receptor function (28).To assess whether this was also true in forebrain neurons, we used live ratiometric calcium imaging via FURA-2AM.When selecting for KCl-responsive cells (neurons), HCMV-infected cells (eGFP + ) demonstrated reduced calcium baselines compared with mock controls (Fig. 10A  and B).This was consistent across all tested lines, regardless of genetic background.Furthermore, infection reduced the average response to KCl stimulation within this neuronal subgroup, highlighting HCMV's potential to reduce calcium mobilization in response to KCl stimulation (Fig. 10C).
Having established that HCMV can alter calcium signaling, we next tested the effect of infection on neuronal electrophysiological function.Cells were plated onto multielec trode arrays and cultured for 50 days (D74 of differentiation).During this period, cells were evaluated for neuronal maturation via action potential generation.Complementing our observations in calcium imaging, electrophysiological activity increased until D54 of differentiation and plateaued, indicating functional maturity (Fig. 10D, denoted by red arrow).Subsequently, each culture was infected with TB40/E-eGFP or mock treated, and recordings were conducted on each plate three times per week.Cultures were main tained in this paradigm for a minimum of 50 dpi (D124 of differentiation).Within 5 days, small punctate regions of eGFP were present within the MEA wells.As indicated by the 7dpi (Fig. 9A and B) and subsequent 14-dpi TUNEL experiments, infection over long periods of time did not result in increased rates of cell death, with mock-and HCMVtreated cells being comparable in the amount of TUNEL-positive nuclei.However, neuronal processes (axons, dendrites) disappeared by 65 dpi in infected cultures (Fig. 10E).Furthermore, long-term HCMV infection effectively eliminated any spontaneous and evoked action potential generation within forebrain neuron cultures, as observed by a lack of action potential firing in the generated Raster plot (Fig. 10F, bottom right subpanel).Time course analysis of spontaneous and evoked activity measured via weighted mean firing rate within cultures highlighted a rapid decline in action potential generation between 0 and 15 dpi across all cell lines (Fig. 10G).The timeline of this decline coincided with the reductions in baseline calcium and KCl response observed via calcium imaging at 7 dpi (Fig. 10A through C, timepoint denoted by red arrows in Fig. 10G).Action potential generation was effectively eliminated after this point in infected cultures, while mock-treated neurons continued to effectively signal.Taken together, these data demonstrate that HCMV infection of mature forebrain neurons disrupts cell calcium signaling and impairs the electrophysiological neural function.

DISCUSSION
HCMV infection is widely known to induce severe developmental neurological deficits due to the clear susceptibility of NPCs to infection observed in vitro and in vivo in the fetal brain (21,22,61,62).More recently, there is also a potential link between HCMV and glioblastoma (63,64) as well as age-related neurodegeneration in Alzheimer's disease (65)(66)(67), which suggest that HCMV may also impact the health and function of terminally differentiated adult neural cells.Here, we show that HCMV robustly infects terminally differentiated iPSC-derived neurons and dramatically disrupts their structure and function.
To test if HCMV could infect terminally differentiated neurons, we differentiated iPSCs from both healthy individuals and patients with Alzheimer's disease in vitro for 84-120 days prior to HCMV infection.Although we do recognize a limitation that iPSC-derived neurons lack full adult in vivo neuron characteristics (68)(69)(70), the neurons used here expressed clear markers of post-mitotic glutamatergic and GABAergic forebrain neurons and showed functionally mature calcium and electrophysiological properties (Fig. 2A, 3 and 10).Action potential generation is a fundamental feature of mature neurons and results from coordinated activity of a variety of voltage-gated ion channels (71).Imma ture iPSC-derived neurons will generally fire a single action potential with low amplitude, whereas mature iPSC-derived neurons will fire rapid chains of action potentials (71).In this regard, we observed iPSC-derived forebrain neurons from all five iPSC lines produced actional potential bursts indicative of neuronal maturation.Previous studies have shown either a decrease in susceptibility to HCMV infection or a more limited impact of HCMV infection as NPCs were pushed toward a more differentiated state (21,72).However, we did not observe this to be the case as neurons in culture for 84-120 days were robustly infected at an MOI of 3 (Fig. 7A, 8G, 9D through F and 10A).The genetic background of the patient cell line also did not impact infection efficiency as all five iPSC lines showed similar results.Additionally, similar data were found in fetal cortex-derived neurons indicating that these results were not simply a feature of the iPSC system.Previous studies analyzing infection of neuronal populations primarily use HCMV strains grown and maintained on fibroblasts (21,22,27).Recent work from our group and others have highlighted differences in HCMV based upon the cell type used for propagation, with divergences occurring in entry receptor expression (73,74), cell tropism (75), and viral spread (74).Therefore, we used epithelial-derived TB40/E-eGFP HCMV due to improved markers of infection relative to fibroblast-derived TB40/E-eGFP HCMV stocks in iPSC-derived neural cultures (74,76).Together, these data demonstrate that HCMV infection is possible even when occurring at points substantially past terminal differentiation.
We and others have shown that NPCs undergo cell death in response to HCMV infection (21,52), but we did not observe significant cell death in infected terminally differentiated neurons (Fig. 9A).However, infected neurons lost characteristic morpho logical and functional features of neurons and showed expression of proliferative markers.Numerous laboratories have demonstrated that HCMV overrides the host cell cycle regulation in order to optimize viral gene expression, viral DNA replication, and viral production (77).In congenital HCMV infection, it is possible that some loss of NPCs is not detrimental to overall viral propagation because of the large NPC pool available in the early developing brain that could be used to maintain optimal viral replication.However, with a finite number of post-mitotic neurons in the adult brain, overt cell loss could impede viral spread particularly if viral replication and/or production is less efficient.It is not known why or how HCMV would have different effects on progeni tor cells compared with terminally differentiated neurons, so future studies assessing transcriptional and proteomic profiles will be needed.
We and others have assessed the impact of HCMV on calcium function in NPCs (28) and other cell types (29,78,79), but the functional impact that HCMV has on termi nally differentiated neurons had not been previously established.Similar to results from NPCs, we found that HCMV dramatically reduced calcium baseline levels and impaired the calcium response to a KCl depolarizing stimulus (Fig. 10A through C), which is likely due to disruptions in voltage-gated ion channel function (28).Action potential generation was effectively eliminated within 2 weeks of infection (Fig. 10G).These data provide important insight into the potential pathological effects of HCMV on the adult brain.For example, HCMV seropositivity is correlated with Alzheimer's disease and a faster rate of cognitive decline (65).Action potentials drive communication between neurons in the CNS, and maintained neuronal connectivity is essential for cognitive health.Therefore, it is possible that HCMV infection in adult neurons significantly reduces neuronal communication, thereby impacting overall cognitive function.More research is needed to directly assess HCMV infection on overall disease pathology, but the data presented here suggest that neuronal HCMV infection could induce significant functional impairment.
In the course of our analyses, a variability was observed between neurons generated from different iPSC lines.While it is imperative to note that these differences did not affect permissiveness or virion production to any substantial degree, potential new avenues of research were highlighted by these altered phenotypes.Specifically, the differences in viral GFP production by the sAD line (Fig. 2C) and mixed effects of HCMV on cell cycle targets (Fig. 8B, C, E, and F) raise questions about how different genetic backgrounds may contribute to altered cellular dynamics in response to infection.Additionally, differential electrophysiological activity demonstrated by the familial AD line PSEN2 N141I could hint at a disease-specific effect on mature neuronal activity both with and without HCMV infection (Fig. 10G).However, these differences are not unexpected in iPSC model systems due to the genetic variability between the patients from which these cells were derived, but more work is needed to better understand genetic diversity and susceptibility to HCMV infection.Finally, although our analysis focused on neurons, we acknowledge that the cultures are not pure and that they contain other glial cells.Astrocytes have been shown to be infected by HCMV (16), so it is possible that the initial infection targets astrocytes rather than neurons.However, neurons in the cultures become infected and show devastating functional and cellular consequences.More work is needed to determine the primary target of infection, but as the brain is not a mono-cellular structure, even if a non-neuronal cell type initiates the infection, our data clearly demonstrate that neurons suffer from the infection.
Upon observing infection in terminally differentiated neurons, one question we had was whether the infection progressed through full viral replication and if infected neurons could produce infectious virus.To answer the first question, we used the TB40/ E-eGFP/mCh dual-labeled virus in which mCherry is driven by the UL99-produced late viral protein pp28.Importantly, pp28 protein is only synthesized after the onset of viral replication (80).Using this viral construct, we observed a steady rise in pp28-mCherry expression in terminally differentiated neurons across all samples beginning around 100 hpi (Fig. 7D).Interestingly, we found mCherry expression in the neurites of infected cells (Fig. 6C) consistent with neuronal transport observed with other neurotropic herpesvi ruses (81)(82)(83).This could indicate a potential cell-cell-mediated mechanism for viral propagation in the adult brain to sites away from the initial infection.This idea is further supported by Silva et al. who demonstrated cell-to-cell spread even in the absence of full virion assembly (84).Moreover, we found robust mCherry expression in structures reminiscent of virion assembly compartments (Fig. 9D) surrounded by areas of organized multi-nucleated syncytia-like structures (Fig. 9D through F).Syncytia formation is widely observed in HCMV-infected tissues and is thought to be involved in viral spread to neighboring cells (60,76,85).The structural changes associated with HCMV infection in terminally differentiated neurons may also play a role in viral propagation in the adult brain.
To assess whether infected neurons could produce infectious virus, we collected the conditioned medium and measured viral titers on ARPE19 epithelial cells (Fig. 7E).Previous studies using NPC cultures have found that infected NPCs can propagate the infection to other cells in the culture (20), but others have found that infected NPCs, even infected at an MOI of 20, produced poorly infectious virions (72).We observed that iPSC-derived neurons from all five lines produced highly infectious virions ( and F), but we did note that neurons generated more non-infectious viral particles compared with fibroblasts (86) (Fig. 7F).It is not clear why this is the case, but there may be differences in how HCMV is packaged and released by neurons compared with fibroblasts.Of note, fetal-derived neurons infected with TB40/E-eGFP were more variable with regard to their produced titers.These apparent differences when compared with iPSC-derived models may be due to a variety of factors.First, while the fetal-derived neuronal cultures utilized in the shown experiments were similar in cell type composition (Fig. 2B and  D), the timepoint at which the fetal tissue was collected was not consistent.Both this potential difference in pre-harvest gestational maturity and variable passage numbers in culture could affect susceptibility to HCMV and production of virions.It should also be noted that each fetal cell line, as is the case with the iPSC lines, has a distinct genetic background, all of which could increase or decrease susceptibility to HCMV infection and affect dynamics of the viral lifecycle.Finally, due to the shorter viability of fetal-derived neurons in culture (compared with iPSC derived), infections in fetal-derived cells were only carried out for 7 days, versus 14 days in iPSC-derived cells.Nevertheless, it will be interesting to further test whether the non-functional viral particles still have an impact on neuronal health and function through the release and expression of tegument proteins.
Together, our studies show that terminally differentiated, functionally mature human neurons are robustly infected by HCMV, which results in a substantial decline in functional properties.Although the detrimental effect of congenital HCMV infection has been well established, the data presented here suggest that HCMV infection in terminally differentiated neurons is similarly detrimental and could impact the overall brain health in the adult population.).Control stem cell lines were chosen based on the availability and diversity of subjects.Controls 1 (21 years old) and 2 (age unknown) are derived from white female patients, while Control 3 (28 years old) is from a male, African American donor.AD lines were also chosen based on the availability and diversity of disease origin/progression.AG25370 (PSEN2 N141I ) originates from a familial AD patient and contains a mutation associated to the amyloid beta processing pathway.AG27607 (sAD) is derived from a sporadic AD patient and possesses no known mutations associated with AD.These lines are both from white females, aged 69 and 81, respectively.No previous testing for HCMV positivity was indicated, and we have not analyzed these samples for markers of latent infection.Stem cells were maintained under feeder-free conditions on Geltrex-coated (Gibco) six-well plates, with daily replacements of Essential 8 Medium (Thermo Fisher Scientific).All iPSCs were grown for a minimum of three passages post-thaw prior to differentiation and were confirmed mycoplasma negative.

Cell culture and differentiation
Patterned differentiation of iPSCs toward neural progenitor cells was conducted using dual-SMAD inhibition (STEMdiff SMADi Neural Induction Kit, STEMCELL Technolo gies).To promote complete and uniform generation of NPCs, cells were maintained in SMADi medium (with daily media changes) for a minimum of three passages prior to further differentiation to neurons.During each passaging step, NPC monolayers were dissociated to single-cell solutions using Accutase solution (STEMCELL Technologies) and subsequently replated at a density of 2 × 10 5 cells/cm 2 onto Geltrex-coated six-well plates.Cultures were maintained for 6 days between passages.At day 18 of differentiation, NPCs were again dissociated and plated at a density of 1.25 × 10 5 cells/cm 2 .
Fetal-derived neuronal cultures were differentiated from neurospheres, as described by Ebert et al. (38).Briefly, ~8-12-week post-conception cortical tissues (M045 and G010) were originally obtained from the University of Washington Birth Defects Research laboratory under an approved Institutional Review Board (IRB) protocol (87,88).Tissue samples were dissociated, cultured as floating aggregates, and passaged by mechan ical chopping for ~20 weeks in DMEM/F12 medium supplemented with 20 ng/mL EGF, 20 ng/mL FGF, and 2% B27.Neurospheres were aliquoted and frozen for later use.To induce differentiation, neurospheres were thawed, dissociated, and plated onto laminin-coated coverslips in minimal Neurobasal medium supplemented with 2% B27.The in vitro use of human fetal tissue was approved by the Medical College of Wisconsin IRB (PRO00025822).

Viruses
HCMV viruses TB40/E-eGFP (39,40) and dual-fluorescently tagged TB40/E expressing IE2-2A-eGFP and UL99-mCherry, generously provided by Felicia Goodrum (University of Arizona, Tucson, AZ) and Eain Murphy (SUNY Upstate Medical University, Syracuse, NY), respectively, originated from transfection of both a bacterial artificial chromosome (BAC) encoding each HCMV subvariant and a UL82-encoding plasmid into MRC-5 fibroblasts.This was accomplished using electroporation (260 mV for 30 ms; 4-mm-gap cuvette) via a Gene Pulser XCell System (Bio-Rad).Subsequently, generated stocks of each fibroblast-derived virus were used to infect ARPE-19 cells, generating epithelial-derived TB40/ E-eGFP and TB40/E-eGFP/mCh.Upon collection, medium containing virus was processed by centrifugation (Sorvall WX-90 Ultracentrifuge and SureSpin 630 rotor; Thermo Fisher Scientific), using a sorbitol cushion (20% sorbitol, 50 mM Tris-HCl [pH 7.2], 1 mM MgCl 2 ) at 20,000 × g for 1 hour.Stock titers were determined by inoculating ARPE19 cells in a 96-well plate and conducting a limiting dilution assay.At 2 weeks post-infection, wells were assessed for either GFP + cells or IE1 + staining, and infectious units were established in infectious units per milliliter.Infections were conducted at an MOI of 3 unless otherwise stated.Infections were conducted by inoculating virus-containing (or virus-free for mock) media to cells for 2 hours, with constant agitation.After 2 hours, media (+/− virus) were removed from each well, cells were washed 1× with Dulbecco's phosphate-buffered saline (PBS, Gibco), and fresh media were added.

Viral titers and infectivity
Viral titers were assessed by collecting conditioned media from infected cultures at 14 dpi (iPSC-derived forebrain neurons), 7 dpi (fetal-derived neurons), and 4 dpi (neural progenitor cells).These timepoints varied due to culture viability.Serial dilutions of conditioned media were administered to plated ARPE-19 cells in a 12-well dish (in duplicate), and the cells were allowed to incubate for 1 week.After incubation, cells were stained using an antibody to viral protein immediate early 1 (IE1, added 1:500), and the IE1 + cells were quantified using Goat α-mouse AF488 (1:1,000).Results were reported as infectious units per milliliter.
DNA was isolated from the same neuronal and NPC-conditioned medium samples as used in the infectious unit assay.One hundred microliters of each sample was pretreated with 2 µL of TURBO DNase (2 U/µL; Invitrogen) and incubated at 37°C for 30 minutes.Subsequently, samples were treated with 1 µL of proteinase K (20 mg/mL), 4 µL of 20% SDS, and 400 µL TEN (Tris, EDTA, and NaCl), vortexed, and incubated overnight at 37°C.DNA was extracted using a phenol-chloroform isolation method.Rendered DNA was then assessed for the number of viral DNA copies using primers for viral gene UL123 via quantitative PCR (qPCR).Primer sets specific to UL123 were utilized (5′-GCCTTCCC TAAGACCACCAAT-3′ and 5′-ATTTTCTGGGCATAAGCCATAATC-3′).Subsequently, the viral DNA copy number from released particles was divided by the titer obtained from the infectious unit assay.The resulting value addresses the infectivity of viral particles at 14 dpi.

Protein and DNA analyses
Evaluation of relative viral-to-cellular DNA levels was completed using qPCR.Cells were collected for timepoints ranging from 1 to 5 days post-infection, with DNA from each being isolated via the DNeasy blood and tissue kit (Qiagen).Subsequently, primer sets for HCMV UL123 (5′-GCCTTCCCTAAGACCACCAAT-3′ and 5′-ATTTTCTGGGCATAAGCCATAA TC-3′) and cellular TP53 (5′-TGTTCAAGACAGAAGGGCCTGACT-3′ and 5′-AAAGCAAATGGA AGTCCTGGGTGC-3′) (Integrated DNA Technologies) were used to determine relative viral DNA over the time course.2× SYBR green PCR Master Mix (Bio-Rad, Thermo Fisher) was utilized for the qPCR.Data collection and analysis were completed using a QuantStudio 6 Flex Real-Time PCR Machine (Thermo Fisher).Results utilized relative HCMV UL123 quantities being normalized to relative cellular TP53 quantities.

Immunofluorescence and live imaging
iPSC-and fetal-derived neurons were plated onto poly-L-lysine/laminin-coated (MilliporeSigma) coverslips at a density of 30-35,000 per coverslip.Cells were fixed via incubation in 4% paraformaldehyde for 15 minutes, washed with dPBS twice, and stored in fresh dPBS until use.Application of Triton X-100 solution (0.2% [vol/vol], in PBS) for 10 minutes permeabilized cells.After washing once with PBS, blocking buffer (5% normal donkey serum [NDS], 5% normal goat serum [NGS], vol:vol, in PBS] was applied to the cells for 1 hour at room temperature.Coverslips were incubated in primary antibody buffer (primary antibody, 2.5% NGS, 2.5% NDS, 0.1% Triton X-100, in PBS) overnight at 4°C and secondary antibody solution (secondary antibody, 2.5% NGS, 2.5% NDS, 0.1% Triton X-100, in PBS) for 1 hour at room temperature.Nuclei were stained using Hoechst 33342 (1:1,000 in PBS; Invitrogen).Coverslips were mounted on glass slides using Fluoromount-G Mounting Medium (Invitrogen) and sealed using clearcoat nail polish.Slides were imaged using a Zeiss LSM980 confocal microscope, with image analysis being conducted in Nikon Elements and Zeiss ZEN analysis software.
Live cell visualization of the virally encoded eGFP and mCherry fluorophores was completed using a Nikon TS100 inverted microscope.Timelapse live imaging was completed using an IncuCyte S3 in-incubator imaging system (Essen Biosciences).Brightfield, green fluorescent (ex.440-480 nm; em.504-544 nm), and red fluorescent (565-605 nm; em.625-705) images were captured every 2 hours for 7 days using a 10× objective.IncuCyte 2022A software was used for background subtraction, image analysis, and video stitching.

Calcium imaging and microelectrode array
Immature neurons were plated onto poly-L-lysine/laminin-coated (MilliporeSigma) coverslips at a density of 30,000 cells per coverslip.After D77 of differentiation, cells were infected with HCMV or mock treated.At 7 dpi, cells were removed from neuronal maturation media, washed once with extracellular normal HEPES (ENH) solution (150 µM NaCl, 10 µM HEPES, 8 µM glucose, 5.6 µM KCl, 2 µM CaCl 2 , 1 µM MgCl 2 ), and loaded with FURA-2AM calcium dye solution (2.5 µM FURA-2AM, 2% [wt/vol] BSA, ENH) for 1 hour at room temperature.Coverslips were then washed again with ENH for 15 minutes before being mounted to a perfusion chamber.Prior to calcium recordings, both green fluorescent and brightfield images were obtained.To establish baseline readings, coverslips were superfused with ENH solution at a rate of 6 mL per minute for 1 minute prior to initial stimulation.Ten micromolars ATP was applied to the cells for 1 minute to elicit a purinergic response, with a 1 minute ENH washout period immediately following.Then, KCl was administered to the cultures for 30 seconds to examine voltage-gated ion channel activity.Imaging ended with a 30-second washout period.A Nikon Eclipse Ti-U inverted microscope was used for imaging, and NIS-Elements Advanced Research (Nikon) was utilized for analysis and image processing.Within each field of view, 50 cells were selected for recording.This experiment was repeated using three regions per coverslip, across two coverslips, per condition.Plotted R 340/380 ratios represent bound (340 nm) to unbound (380) calcium at baseline.
Cells were plated onto a PLL/laminin-coated (MilliporeSigma) CytoView microelec trode (MEA) plate (Axion Biosystems) at a density of 30,000 per well.Cells were recorded three times weekly using a Maestro MEA system (Axion Biosystems), analyzing both spontaneous and electrically evoked action potentials.For spontaneous recordings, cells were recorded for a 6-minute window with the default settings for spontaneous recordings set within the AxIS software suite (Axion Biosystems).Analysis of spontaneous data was conducted using the Neural Metric Tool (NMT, Axion Biosystems).In NMT, the criterion for an "active electrode" was defined as 5 spikes/minute, and coincidence artifacts were removed.Additionally, in generating time course data, wells that never demonstrated activity were removed from any subsequent analysis.Data are reported as weighted mean firing rate.Electrically evoked potentials were recorded over a dura tion of 2 minutes (following spontaneous recordings), with 0.5 V stimulations (400-µs duration, each) occurring every 10 seconds.As with spontaneous recordings, evoked recordings were processed using NMT to determine active electrodes and remove coincidence artifacts.Additionally, any spikes occurring within 2 ms of the electrical stimulation were discarded to ensure detected action potentials were not artifactual.
Neuron-free wells produced evoked wMFR values of around 0.15 Hz, leading to a background subtraction of 0.15 Hz from all wMFR values generated by NMT (evoked data only).As with spontaneous activity data, wells never exceeding this 0.15 Hz threshold were removed from further analyses.

Funder Grant(s) Author(s)
Stead Family Foundation Scott S. Terhune

FIG 1 4 FIG 2
FIG 1 Assessed neurons were generated from both iPSC-and fetal-derived progenitor cells.(A) iPSC colonies were dissociated, plated into SMADi Neural Progenitor Medium, and differentiated into neural progenitor cells over an 18-day period.Generated NPCs were dissociated/plated into forebrain differentiation medium for 6 days to prepare immature forebrain neurons.Final plating of immature neurons occurred at D24 of differentiation, and cells were maintained in forebrain neuron maturation medium until use.Timepoints of HCMV infection are denoted by green virus icons, and culture collections are indicated by red octagons.(B) Fetal-derived cultures were collected from forebrain tissue at ~D63 of in utero development (~9 weeks).Neural progenitor samples were cultured as neurospheres and expanded for an average of 133 days (19 weeks) prior to neuronal differentiation.Neurospheres were differentiated over a 1-week period and plated down at ~D203 of differentiation.Plated neurons were allowed to grow for 23 days prior to HCMV exposure (D226), with infection persisting for an additional 7 days (D233) prior to sample collection.Scale bars = 200 µm.

FIG 2 ( 6 FIG 4
FIG 2 (Continued) neuronal cells and 37.39% glial cells (including astrocytes and progenitors).(C) Similar neuronal populations were found in fetal progenitor-derived cultures, with both excitatory and inhibitor neurons being present.(D) Similar to iPSC-derived cultures, fetal-derived neurons contained 64.81% neurons and 35.19% glial cells.Data are presented as mean ± SEM.Scale bars = 50 µm.

FIG 6 10 FIG 7
FIG 6 Fetal-derived neurons demonstrate HCMV spread.Infection of day 23 fetal-derived neurons with TB40/E-eGFP reveals progressive increases in fluorescent signal starting as early as 4 dpi.Large eGFP + plaques are visible by 10 dpi, accompanied by a significant alteration to cell morphology.Scale bars = 100 µm.

FIG 7 ( 12 FIG 8
FIG7 (Continued)    fluorescence measurements highlight increases in both eGFP and mCherry throughout the initial 7 days after HCMV exposure (n = 3).(E) Assessment of viral titers within 14 dpi neuron-conditioned media (inoculum) was completed using ARPE-19 epithelial cells.All tested forebrain neuron cultures produced relatively consistent viral titers regardless of the timepoint being assessed.These values are comparable to those seen in tested neural progenitors at 4 dpi.Fetal-derived neurons (7 dpi) demonstrated inconsistent results, with one of two lines producing viral titers (n = 2-4).(F) Infectivity of viral products was assessed using viral DNA copies present within neuron-conditioned media relative to that sample's viral titer, shown in E. As with the titer values, infectivity was consistent across all forebrain neuron lines and timepoints.Furthermore, forebrain neurons values were consistent with NPC infectivity ratios.As in E, fetal-derived cells produced highly variable infectivity.All data are presented as mean ± SEM.Two-way ANOVA was used to analyze variance in panel B and among iPSC cultures in E and F. *P < 0.05, **P < 0.01, and ***P < 0.001.

FIG 9 ( 15 FIG 10
FIG9 (Continued)    revealed the presence of pp28-mCherry signal within neuronal processes, indicating a potential trafficking mechanism (indicated with blue arrows).(D) Using TB40/E-eGFP/mCh, pp28 signal can also be observed organizing into virion assembly compartments surrounded by rings of nuclei (syncytia).(E) These findings are consistent with pp28 immunofluorescent staining patterns.(F) Syncytial structures are formed in both iPSC-and fetal-derived cultures.Furthermore, Tuj1 positivity indicates that neurons are the components of the observed syncytia.

FIG 10 (
FIG 10 (Continued) calcium are significant across all tested lines (n = 188-296 cells per condition).(C) Upon stimulation with KCl, HCMV dampens calcium influx (n = 188-296 cells per condition).(D) Multielectrode array (MEA) recordings of forebrain neuron electrical activity from the date of plating until TB40/E-eGFP infection.Neuronal activity increases until D50 of differentiation, indicating functional maturity (denoted by red arrow).Due to aberrant electrophysiological activity, Control 1 was excluded from subsequent analyses.(E) Representative images of mock-treated and infected forebrain neurons plated onto MEAs at 65 dpi.(F) Representative raster plots showing electrical activity at 5 and 65 dpi.Spontaneous action potentials are nearly eliminated in HCMV cultures over time, whereas mock-treated cells still retain functional capabilities.(G) HCMV's effects on action potential generation (both spontaneous and evoked) are present in as early as 7 dpi, correlating with calcium imaging findings in A-C (timepoint indicated by red arrow).By 20 dpi, most electrical activity has ceased.This finding is ubiquitous across all tested lines.Analyses are restricted to MEA wells that demonstrate activity above threshold (wMFR > 0.15) (n = 5-15 wells per condition).Statistics in panels B, C, and G were completed using two-way ANOVA.†P < 0.1, *P < 0.05, **P < 0.01, and ***P < 0.001.