Neutralizing antibodies with neurotropic factor treatment maintain neurodevelopmental gene expression upon exposure to human cytomegalovirus

ABSTRACT

Human cytomegalovirus (HCMV) is a beta herpesvirus that causes severe congenital birth defects including microcephaly, vision loss, and hearing loss. Entry of HCMV into human cells is determined by the composition of glycoproteins in viral particles and influenced by the source of the virus. We have previously shown that HCMV infection of human stem cell-derived cerebral organoids causes significant downregulation of critical neurodevelopmental genes, but the mechanism of viral entry in human neural tissues is not well known. Here, we evaluated infection efficiency using virus from different sources. We observed significant increases in the number of viral genomes, viral spread and penetrance, and multinucleated syncytia in neural tissues infected with HCMV propagated in epithelial cells compared with fibroblasts. Interestingly, we found similar expression levels of cellular entry receptors on organoid-derived cells suggesting that neural cells do not exhibit a biased entry receptor expression profile. Next, we asked whether we could limit viral entry using neutralization antibodies. We found that pre-treatment with antibodies against viral glycoproteins gB and gH successfully decreased viral genome levels, viral gene expression, and virus-induced syncytia. In contrast, targeting specific cellular entry receptors failed to limit infection. Using an antibody against gB, we observed partial protection of neurodevelopmental gene expression that was further improved by the addition of brain-derived neurotrophic factor (BDNF). These studies indicate that the source of HCMV is a key determinant of viral entry into neuronal cells and combining gB neutralization with BDNF provides further benefit to neural gene expression during infection.

IMPORTANCE

Human cytomegalovirus (HCMV) infection is the leading cause of non-heritable birth defects worldwide. HCMV readily infects the early progenitor cell population of the developing brain, and we have found that infection leads to significantly downregulated expression of key neurodevelopmental transcripts. Currently, there are no approved therapies to prevent or mitigate the effects of congenital HCMV infection. Therefore, we used human-induced pluripotent stem cell-derived organoids and neural progenitor cells to elucidate the glycoproteins and receptors used in the viral entry process and whether antibody neutralization was sufficient to block viral entry and prevent disruption of neurodevelopmental gene expression. We found that blocking viral entry alone was insufficient to maintain the expression of key neurodevelopmental genes, but neutralization combined with neurotrophic factor treatment provided robust protection. Together, these studies offer novel insight into mechanisms of HCMV infection in neural tissues, which may aid future therapeutic development.

INTRODUCTION

Human cytomegalovirus infection is the leading cause of non-heritable birth defects worldwide and remains a serious threat of morbidity for immunocompromised individuals (1, 2). Although inhibitors of viral DNA replication are available, adverse side effects limit their use in pregnant people (3), thereby limiting adequate treatment options for this highly vulnerable patient population. Congenital human cytomegalovirus (HCMV) infection occurs when the virus is passed from the pregnant person to the fetus. Between 2% and 6% of neonates are diagnosed with congenital HCMV infection around the world each year, and it is the most common congenital infection in the United States impacting nearly 30,000 babies each year (1, 4, 5). The most serious cases cause lifelong neurological abnormalities such as microcephaly, sensorineural hearing loss, and developmental motor delays (4, 6, 7). Worldwide, 45%–100% of individuals is estimated to be seropositive in the general population (5, 8, 9). Fortunately, symptoms are typically mild for a healthy individual, and the virus will eventually enter a latent stage with occasional reactivations remaining in the body (4, 8). This life-long persistence has been associated with diverse diseases such as immuno-senescence, atherosclerosis, gliomas, and Alzheimer’s disease (8, 10).

One reason for HCMV’s high prevalence is its broad tropism infecting many different human cell types including endothelial cells, epithelial cells, fibroblasts, lymphocytes, macrophages, and neural cells (11, 12). Until recently, most studies characterizing viral infection were conducted using non-neuronal model systems, leading to questions regarding entry into neural cells and the mechanisms by which the virus disrupts normal neurodevelopmental pathways (12 – 16). What is clear is that HCMV infection in human neural progenitor cells and cerebral organoids causes massive transcriptional downregulation of several developmentally critical transcription factors, dysregulation of calcium signaling and action potential generation, and disruption of cortical layer development and neural rosette formation (6, 9, 11, 17 – 27). For example, Brown et al. (17) report decreases in organoid size, number of developmental rosette sites, and cortical layer depth in cerebral organoids following HCMV infection. Sun et al. (27) have shown that neutralizing antibodies or small interfering ribonucleic acid (siRNA) knockdown of specific host receptors can limit viral entry of HCMV into cerebral organoids. These treatments also correct the decrease in organoid size and cortical layer depth phenotypes and can improve previously noted calcium signaling and action potential defects in HCMV-infected organoids (22). However, it remains unclear what specific viral genes or proteins cause the neurodevelopmental abnormalities or whether viral entry is a necessary part of the mechanism. Our previous work showed that full viral replication was not required to induce transcriptional and functional abnormalities in neural tissues and that downregulation of key developmental genes was not fully dependent on immediate-early IE1 and IE2 expression (18). Here, we focused on better understanding what role viral entry plays in neurodevelopmental changes.

HCMV entry into target cells requires several viral glycoproteins, which include gH/gL and gB. The viral glycoprotein gH/gL complex binds to host cell surface receptors whereas gB promotes fusion and pH-dependent entry (28, 29). Further, the gH/gL glycoprotein complex can form a disulfide-linked trimer with glycoprotein O (gO) to make the trimeric complex gH/gL/gO (TC) (12, 30 – 32) or it can interact with viral proteins UL128, 130, and 131 to form the pentameric complex gH/gL/UL128-131 (PC) (28, 31, 32). A recent unbiased screen of potential HCMV viral receptors identified a list of host receptors that interact with the TC (NRG2, PDGFRα, and TGFβRIII) or PC (CD46, Nrp2, FCAR, and LILRB3) (13, 16, 28, 33, 34). TC interactions are reported to be required for entry into fibroblasts, and previous publications indicate an interaction with the PDGFRα receptor is common and may be required for fibroblast entry (13, 34). Entry into epithelial, endothelial, or myeloid cells requires both TC and PC interactions, and several different potential interacting receptors have been reported or identified (12, 28). These studies indicate that HCMV utilizes different envelope glycoproteins to engage functionally unrelated entry receptors to provide redundancy for viral entry (28, 31). In epithelial cell entry, both TC and PC interactions are required indicating that a viral particle only expressing one of these complexes would fail to produce a successful infection in epithelial cells (12, 28). Viral stocks produced in the lab with continuous passaging on a single cell type might bias the particle to infect that cell type. Specifically, recent evidence suggests that over time, viral particles passaged on fibroblasts will accumulate a higher expression of the TC whereas those passaged on epithelial cells will maintain a more balanced expression of both TC and PC (35).

In the studies described here, we focus our attention on the influence of HCMV entry in neural infection and changes in viral and cellular gene expression. We evaluate the contribution to HCMV entry using virus from a common genetic background yet produced on different cell types for their ability to replicate and spread in neuronal tissues. Further, we assess entry receptor expressions to better understand HCMV tropism for neuronal tissues. Finally, we test the effects of neutralizing antibodies and the neurotrophic factor brain-derived neurotrophic factor (BDNF) on protecting neurodevelopmental gene expression from HCMV infection-mediated damage.

RESULTS

HCMV propagated on epithelial cells exhibits increased tropism for cerebral organoids

Previous studies have found that HCMV propagated on different cell types influences tropisms likely due to changes in virion glycoproteins and available host entry receptors (12, 28, 31, 35 – 37). It is unknown how this impacts infection in neural tissues. To begin addressing this point, we generated human-induced pluripotent stem cell (iPSC)-derived cerebral organoids from a healthy control iPSC line (38, 39) and infected after 30 days (30d) of development using a BAC-derived recombinant HCMV strain TB40/E expressing eGFP (40). Day-30 organoids consist of several cell types including neural progenitors, radial glia, and early born neurons capable of action potential firing, and these tissues contain structural features observed in 9–12 weeks of human fetal brain development (38, 39, 41 – 43). Viral stocks were generated on either epithelial cells, referred to as TB40-BAC4_epi, or fibroblasts, referred to as TB40-BAC4_fib. We infected organoids using a multiplicity of infection (MOI) of 500 infectious units per µg of tissue (IU/µg) due to the natural variance in size and cell number in developing organoids. During infection, organoids were placed on a rocker and exposed to the inoculum for 2 h and then washed with PBS. Infection was allowed to progress for 14 days post infection (dpi). We observed GFP fluorescence by 4 dpi, and this signal continued to increase over the subsequent 10 days indicating viral spread through the tissue (Fig. 1A and B). Tissue sizes remained constant for the duration of the experiment regardless of conditions. Fluorescence was substantially higher in organoids infected with TB40-BAC4_epi compared with TB40-BAC4_fib, despite originating from a common genetic background. To quantify differences, we measured viral genomes relative to a cellular gene using qPCR observing approximately threefold more genomes upon infection by TB40-BAC4_epi versus TB40-BAC4_fib by 3 dpi (Fig. 1C). We also determined the mean fluorescence intensity of projection z-stacks at 14 dpi revealing 10-fold higher fluorescence when infected using TB40-BAC4_epi compared with TB40-BAC_fib (Fig. 1D). These data demonstrate that virus generated from epithelial cells exhibits increased tropism for neuronal tissues and is consistent with past studies investigating PC-mediated viral entry (12, 35, 44).

Fig 1.

HCMV propagated on epithelial cells exhibits increased tropism for cerebral organoids. Infection of iPSC-derived cerebral organoids at 30 days (30d) of differentiation with HCMV at an multiplicity of infeciton (MOI) of 1 IU/µg using virus produced from MRC5 fibroblasts (TB40-BAC4_fib) or ARPE19 epithelial cells (TB40-BAC4_epi). (A) Overlayed brightfield and GFP images of three representative (30d) organoids infected with TB40-BAC4_epi at 4, 6, 8, 10, 12, and 14 days post infection (dpi) show considerable spread and propagation of GFP signal. Size bar 200 µm. (B) Overlayed brightfield and GFP images of three representative 30d organoids infected with TB40-BAC4_Fib at the same time points show the spread and propagation of GFP signal. (C) Viral genomes measured by quantative polymerase chain reaction (qPCR) using primers to the UL123 gene relative to GAPDH and whole organoid DNA infected with TB40-BAC4_fib or TB40-BAC4_epi. (D) Average intensity of GFP fluorescence within Z-stack images of TB40-BAC4fib- or TB40-BAC4_epi-infected organoids at 14 dpi. All images were captured at 5× using a Z-stack imaging protocol on a Zeiss LSM9800 microscope. In panel (C) for TB40-BAC4_fib, n = 4, and for TB40-BAC4_epi, n = 4 including all those shown in panel (A). In panel (D) for TB40-BAC4_fib, n = 6, and for TB40-BAC4_epi, n = 6 including all those shown in panel (A). Significance was determined using Student’s t-test. *P < 0.05, **P < 0.01, and ***P < 0.001.

We quantify infection in a three-dimensional (3D) space to assess viral spread and depth of penetrance for each viral source. We rendered the 2D z-stack images taken at 14 dpi into 3D using Zeiss Zen Blue image processing software (Fig. 2A and B). We have included maximum-intensity projection images with heat maps to indicate the depth of signal. Infection using TB40-BAC4_epi reached an average depth of 650 µm from the organoid surfaces of four organoids compared with TB40-BAC4_fib at an average depth of 525 µm (Fig. 2C). Penetrance was further quantified by measuring the GFP signal intensity at the centermost z-plane from stacks imaged at 14 dpi, which shows significantly higher GFP signal at center slices of TB40-BAC4_epi-infected tissues (Fig. 2D). Finally, we analyzed differences in the 3D space by binning GFP signal intensities into 25 equal bins. We find that organoids infected by TB40-BAC4_epi have a higher frequency of voxels (i.e., 3D pixels) within each GFP intensity bin versus TB40-BAC4_fib-infected organoids (Fig. 2E). As such, infection with TB40-BAC4_epi resulted in a higher frequency of GFP signal across all dimensions of the organoid when compared with organoids infected with TB40-BAC4_fib. Additionally, the difference in the frequency of voxels within each bin was largest for bins 0–5 containing the highest GFP intensities indicating that a more severe infection results from the use of TB40-BAC4_epi. Together, these data demonstrate that epithelial-derived virus exhibits greater tropism for neural tissues than fibroblast-derived virus, impacting viral spread and the depth to which the infection can penetrate. To control for any errors relating to the inoculum, we re-titered stocks on ARPE19 epithelial cells, MRC5 fibroblasts, and iPSC-derived neural progenitor cells (NPCs). Titers for TB40-BAC4_fib were 5.4 × 10⁸ IU/mL on MRC5 and approximately 1 log lower at 5.7 × 10⁷ IU/mL on ARPE19 and 5.8 × 10⁷ IU/mL on NPCs. In contrast, TB40-BAC4_epi virus showed similar titers on all three cell types at 2.3 × 10⁸ IU/mL on MRC5s, 1.8 × 10⁸ IU/mL on ARPE19s, and 2.4 × 10⁸ IU/mL on NPCs. Consistent with studies by others (35), these findings support the conclusion that HCMV, originating from the same genetic background of TB40-BAC4, exhibits increased tropism for cortical tissues when originating from epithelial cells. It’s important to note that infection by virus produced from fibroblasts still occurred.

Fig 2.

Infection by HCMV TB40-BAC4_epi results in greater spread and depth of penetrance compared with TB40-BAC4_fib. (A) Max intensity projection images along with heat maps of GFP penetrance from two representative TB40-BAC4_epi-infected 30d organoids at 1 IU/µg are wide spread across the surface and penetrance of up to 645 µm. (B) Max intensity projection images along with heat maps of GFP penetrance from two representative TB40-BAC4_fib-infected organoids with penetrance of up to 465 µm. All images were captured at 20× using a Z-stack imaging protocol on a Zeiss LSM9800 microscope. (C) Binning of the GFP average intensity projection image based on signal intensity (ascending from 1 to 25) shows TB40-BAC4_epi has an increase in the frequency of voxels falling into the highest intensity bins compared with the TB40-BAC4_fib. Six 3D organoid images were analyzed per group at 14 dpi using organoids from two differentiations. Each data point is an average of the frequency of voxels from each of the 6 TB40-BAC4fib or TB40-BAC4epi organoids within each GFP intensity bin. (D) A measure of the GFP signal at the most central z-plane of TB40-BAC4_epi- versus TB40-BAC4_fib-infected organoids (n = 6). (E) Quantification of deepest depth of penetrance based on rendered images in (A) and (B) and four additional organoids in each group (n = 6). Significance was determined using Student’s t-test. *P < 0.05 and ****P < 0.0001

NPCs and cerebral organoids express receptors for both trimeric and pentameric HCMV complexes

After observing the increases in infection by TB40-BAC4_epi compared with TB40-BAC4_fib, we hypothesized that cells within the organoids might express higher levels of receptors for the viral pentamer complex (Nrp2, THBD, and CD46) compared with those for the trimer complex (PDGFRα and TGFβRIII) (Fig. 3A). Others have demonstrated the importance of PDGFRα in HCMV infection in stem cell-derived cortical organoids (27). Analysis of RNA levels from our recently published studies using bulk RNA-seq of uninfected organoids (18) shows expression of Nrp2, CD46, THBD, TGFβRIII, and PDGFRα with substantially higher levels of Nrp2 and CD46 compared with other receptors based on normalized count values of fragments per kilobase of transcript per million mapped reads (FPKM) (Fig. 3B). Here, we began by staining unpermeablized cells for surface expression of PDGFRα, Nrp2, and TGFβRIII along with a membrane marker to validate surface expression of these receptor targets within the NPCs (Fig. 3C). We observed co-localization between the receptor target and membrane staining indicating surface expression with PDGFRα and Nrp2 exhibiting stronger signals than TGFβRIII.

Fig 3.

NPCs and cerebral organoids express several receptors required for HCMV entry at the plasma membrane. (A) Schematic made in BioRender depicting the HCMV viral particle noting the trimeric (gHgLgO) and pentameric (gHgLpUL128-131A) glycoprotein complexes and target cellular receptors PDGFRα and TGFβRIII (trimer) or Nrp2, THBD, and CD46 (pentamer). (B) RNA expression levels of these entry receptors as determined by analyzing previously performed bulk RNA-seq analysis in cerebral organoids (11). (C) Immunostaining conducted in passage 3 NPCs 3 days post plate down labeling receptors (PDGFRα, Nrp2, and TGFβRII) in green, membrane marker (red), Hoescht (blue), and merged to show the co-localization of membrane and receptor targets. This staining was conducted without permeabilizing the cells to strictly focus on surface expression of the receptors.

To quantify surface expression for all receptors, we used flow cytometry on uninfected NPCs and 45d organoids. Dissociated cells from either cultured NPCs or organoids were analyzed for expressions of TGFβRIII (TC), PDGFRα (TC), Nrp2 (PC), CD46 (PC), and THBD (PC) using AF488, APC, or AF405 for which the gating is shown in Fig. 4A. For each sample, 10,000 events were recorded, biological replicates were averaged, and we show the data as a percentage of single or dual positive cells for the indicated receptors (Fig. 4B and C). We observed roughly 50% of cells in the NPC population expresses receptors for the virion trimer with PDGRFα being seen in 10% more cells than TGFβRIII (Fig. 4B). Receptors for the pentamer, Nrp2, CD46, and THBD, were detected in approximately 60% of the population. When evaluating cells expressing receptors for both the HCMV trimer and pentamer complex, we detected similar percentages for PDGFRα-Nrp2 or PDGFRα-CD46, yet reduced percentages for THBD-TGFβRIII receptors. Repeating the analysis using cells from dissociated organoids, we observed similar distributions of the receptors on cells, but at percentages lower than those for NPCs (Fig. 4C). These studies demonstrate that entry receptors for both the trimer and pentamer complexes are ubiquitously expressed in 2D cultures and 3D neural tissues, including evidence of co-expression of receptors.

Fig 4.

Flow cytometry confirms that NPCs and organoids express several receptors required for HCMV entry. (A) Gating strategy for flow cytometry assessing cell surface expression of entry receptors on iPSC-derived NPCs and cerebral organoids. Row 1 from left to right depicts traces from unlabeled cells, gate set for live cells labeled with AF488, APC, and AF405. Row 2 from left to right depicts traces from fluorescence compensation controls for the secondary antibodies. (B) Percentage of NPCs expressing the indicated receptors grouped by preferences or in combination. (C) Percentage of cells isolated from organoids expressing the indicated receptors grouped by preferences or in combination. Results were combined from three separate organoid differentiations and four NPC differentiations with at least two biological replicates per differentiation using one healthy control iPSC line for organoids and two for NPCs. Statistics were determined using Student’s t-test for two groups or ANOVA with Tukey’s post hoc for comparison of three or more groups. *P < 0.05 and **P < 0.01. ns, not significant.

HCMV has been shown to efficiently infect NPCs (11, 19), and recently, Wu et al. demonstrated the importance of SOX2 in regulating HCMV gene expression (24). Therefore, we repeated our analysis to measure cellular entry receptor expression on cells also expressing the neural progenitor marker SOX2. We assessed expression using immunofluorescence from uninfected, plated NPCs derived from SOX2-GFP iPSCs. We stained cells without permeabilization using antibodies against entry receptors PDGFRα, Nrp2, and TGFβRIII and Hoescht (DNA stain) (Fig. 5A). In all samples, SOX2 is colocalizing with the DNA nuclear stain as expected. The entry receptors PDGFRα, TGFβRIII, and Nrp2 exhibit staining on the periphery of cells, indicating cell surface expression (Fig. 5A) as we also show in (Fig. 3C). The slight reduction in number of cells staining for TGFβRIII versus PDGFRα or Nrp2 is consistent with the flow cytometry analysis (Fig. 4). We quantified the percentage of cells exhibiting SOX2 expression in NPCs and organoids using flow cytometry. These experiments were conducted using NPCs without SOX2-GFP and requiring permeabilization followed by staining using SOX2-AF405-conjugated antibody. A higher portion of cells within the NPC cultures, nearly 80%, expressed SOX2 compared with 50% in organoids (Fig. 5B), which is consistent with increased cell type diversity within the tissues. Co-expression of SOX2 with viral entry receptors ranged between 40% and 50% of NPCs (Fig. 5C) with cells expressing PDGFRα, Nrp2, and CD46 being in greater abundance. These percentages were reduced in cells isolated from organoids, which ranged from 30% to 40% of cells (Fig. 5D) reflecting the reduced SOX2 numbers. In summary, HCMV entry receptors for both the viral TC and PC are expressed on both NPCs and cells from organoids, including SOX2-positive progenitor cells. These data suggest that the increased tropism of TB40-BAC4_epi compared with TB40-BAC4_fib is likely unrelated to receptor expressions and may be more related to intrinsic features of the virion and associated entry process.

Fig 5.

Viral entry receptors are expressed on SOX2-positive cells. Panel of staining of passage 3 NPCs 3 days post plate down labeling for (A) receptors (PDGFRα, Nrp2, and TGFβRII) in red, Hoechst (blue), progenitor marker SOX2 (green), and merged. This staining was conducted without permeabilizing the cells to strictly focus on surface expression of the receptors, and the NPCs were derived from iPSCs that express SOX2-GFP. (B) Percentage of cells expressing the progenitor marker SOX2 in NPCs and organoids as determined using flow cytometry. (C) Percentage of cells co-expressing SOX2 with entry receptor in NPCs and organoids. Statistics were determined using Student’s t-test for two groups or ANOVA with Tukey’s post hoc for comparison of three or more groups. **P < 0.01 and ****P < 0.0001. ns, not significant.

Anti-HCMV neutralizing antibodies limit infection of cerebral organoids

Our end goal is to define HCMV neuropathogenesis and identify approaches to limit it using human neural tissue models. To reach this goal and considering our data showing the expression of many possible host entry receptors on NPCs and organoids, we tested the impact of antibodies against viral glycoproteins previously shown to be neutralizing (14, 32, 44). We first pretreated NPCs for 1 h prior to infection with antibodies (4µg/mL) against HCMV gB, gH, or the host receptor PDGFRα. We then infected NPCs with TB40-BAC4_epi at an MOI of 1 IU/cell in the presence of antibody for 2 h. The virus and antibody were then removed, and the cells were washed with PBS and supplied fresh media without antibody. Under these conditions, anti-gB and anti-gH antibodies reduced cell-associated viral genome levels with the largest reduction occurring upon neutralizing gB (Fig. 6A). We did not detect reductions in viral genomes when using an antibody against cell surface receptor PDGFRα (Fig. 6A). Disruption of PDGFRα using siRNA has been shown by others to limit HCMV infection of organoids (27). We evaluated the ability of anti-HCMV antibodies to limit infection per cell using immunofluorescence and staining at 48 hpi. In untreated infections, we observed most cells expressing HCMV IE1 (Fig. 6B) including evidence of multinucleated syncytia with a central cytosolic assembly compartment. These infected cells exhibited enlarged nuclei, which is a hallmark of infection (45), and diffuse SOX2 staining. This was also observed when we tried to block using single antibodies against host receptors like PDGFRα. In contrast, neutralization using either anti-gH or anti-gB antibodies substantially limited the number of GFP- and IE1-positive cells while maintaining nuclear SOX2 (Fig. 6B). Similar results were shown by Sun and colleagues (27) when using an antibody against pUL128/130/131 but completed under different experimental conditions. Finally, we quantified changes in viral gene expression focusing on neutralization with anti-gB antibodies. We measured the expression of the immediate early gene UL123 (IE1 protein), early gene UL44, and late gene UL99. The addition of anti-gB during infection of NPCs reduced expression levels for all gene classes at all time points (Fig. 6C).

Fig 6.

Anti-HCMV neutralizing antibodies limit infection of cerebral organoids. Plated NPCs were pretreated with the indicated antibodies at a concentration of 4 µg/mL and then infected using TB40-BAC4_epi at an MOI of 1 IU/cell. After 2 hpi, cultures were washed with PBS and analyzed at the indicated times. (A) Levels of viral genomes relative to a cellular gene were determined using qPCR on whole-cell DNA isolated at 24, 48, and 72 hpi (n = 4, n = 2 for PDGFRa-treated group). (B) Images of immunofluorescence staining were done at 48 hpi on infected NPC or infected with either anti-PDGFRα anti-gB or anti-gH antibodies. Staining includes TB40-BAC4_epi eGFP (green), anti-IE1 (purple), SOX2 (red), and Hoechst (blue). All images were taken at 40× using the Zeiss LSM9800 microscope. (C) Quantification of viral transcripts was completed using RT-qPCR and primers to UL123, UL44, and UL99 relative to GAPDH (n = 4). Statistical significance was determined using Student’s t-test. *P < 0.05 and **P < 0.01.

Infection by HCMV causes IE1-dependent and independent changes in neurodevelopmental gene expression. In our previous work, we demonstrated that HCMV infection reduces the expression of FOXG1, FEZF2, DMRTA2, and EMX1 in both organoids and NPCs after infection (18). Further, the downregulation of these key gene regulators involved in neurogenesis, cortical layer development, and progenitor maintenance appears to be independent of IE1 and IE2 protein expression during infection (18, 46 – 49). However, it was unclear how these genes would be impacted by neutralizing antibody treatment limiting viral entry. Consistent with our previous results, NPCs infected in the absence of gB antibody resulted in significant reductions of EMX1, FOXG1, DMRTA2, and FEZF2 transcripts compared with mock conditions by 24 or 48 hpi (Fig. 7). In contrast, expression levels were similar to mock upon the addition of the neutralizing antibody at early time points post-infection. While the anti-gB antibody sustained FEZF2, EMX1, FOXG1, and DMRTA2 expression immediately following infection in NPCs, the effects on FEZF2 and FOXG1 were transient and lost by 72 hpi.

Fig 7.

Neutralizing antibodies reduce HCMV-mediated disruption of developmental transcription factors. NPCs were mock-treated, infected using TB40-BAC4_epi at an MOI of 1 IU/cell or infected with the inclusion of an anti-gB antibody at a concentration of 4 µg/mL. Analysis of FOXG1, FEZF2, EMX1, and DMRTA2 gene expression at 24, 48, and 72 hpi. Total RNA was collected, reverse transcribed, and measured using primers to the specific genes. Samples were normalized to GAPDH, and the data are from two iPSC lines across a total of four differentiations, two per line (n = 4). Statistical significance was determined using ANOVA with Tukey’s post hoc comparison. *P < 0.05 and ****P < 0.0001.

We devised an alternative strategy combining the entry-blocking benefits of gB treatment with the neurotrophic factor brain-derived neurotrophic factor. BDNF has been shown to have an important role in cell survival, neurogenesis, and neuronal differentiation (50 – 54), and it is downregulated by murine CMV (55) and during HCMV infection of cerebral organoids (18). We treated mock and HCMV-infected NPCs with either 4µg/mL (high) or 1µg/mL (low) concentration of anti-gB antibody in combination with BDNF at 20ng/mL. Addition of neutralizing antibody at both concentrations significantly reduced HCMV UL123 and UL44 RNA expression at 48 and 72 hpi with a downward trend at 24 hpi (Fig. 8A). A small drop in expression was observed in BDNF-only samples at 72 hpi. These differences were similar to changes seen by GFP fluorescence at 72 hpi (Fig. 8B). Across the neurodevelopmental targets EMX1, FEZF2, and FOXG1, we determined that infection by TB40-BAC4Epi disrupted their expressions as previously observed (Fig. 8C). The addition of BDNF during infection resulted in a small increase in expressions. To our surprise, addition of anti-gB antibody alone in the absence of infection did negatively impact expression of these genes, perhaps by a non-specific association with cells in the absence of virus. When combined, anti-gB antibody and BDNF maintained significantly higher target gene expression in HCMV-infected conditions which occurred at multiple time points post-infection compared with no treatment (Fig. 8C). Antibody neutralization of HCMV infection occurred under all conditions in the population based on GFP fluorescence (Fig. 8B). Together, these data show that the addition of BDNF with anti-gB neutralizing antibodies further limited infection-mediated disruption of several key neurodevelopmental genes in NPCs.

Fig 8.

Neutralizing antibodies with BDNF provide significant protection against HCMV-mediated disruption of developmental factor expressions. NPCs were mock treated, infected using TB40-BAC4_epi at an MOI of 1 IU/cell or infected with the inclusion of an anti-gB antibody at 4 µg/mL (high), 1 µg/mL (low), or BDNF at 2 ng/mL, and analyzed at the indicated time points. (A) Quantification of viral transcripts was completed using RT-qPCR and primers to UL123, UL44, and UL99 relative to GAPDH (n = 4). (B) Merged images of brightfield and GFP fluorescence of live cells taken 72 hpi from the various treatment conditions. (C) Analysis of FOXG1, FEZF2, and EMX1 gene expression at 24, 48, and 72 hpi. Total RNA was collected, reverse transcribed, and measured using primers to the specific genes. Samples were normalized to GAPDH. These results were collected from two WT iPSC lines across a total of four differentiations, two per line (n = 4). Statistical significance was determined using ANOVA with Tukey’s post hoc comparison. *P < 0.05 and **P < 0.01.

DISCUSSION

We and others have shown that HCMV infection induces significant disruption of neural development and function (6, 17 – 19) indicating a need to fully understand the process of infection in human neural tissues. Further, studies by several labs have demonstrated repeated passaging of viral stocks on fibroblasts results in increased accumulation of the trimeric versus pentameric components on the surface of the viral particle (35). In contrast, virus passaged on epithelial cells retains a more balanced composition (35). In our experiments, we evaluate HCMV entry using virus from a common genetic background, TB40E, yet produced on different cell types, MRC-5 fibroblasts or ARPE19 epithelial cells, for their ability to replicate and spread in neuronal tissues. We found that infection using BAC-derived epithelial-propagated virus TB40-BAC4_epi resulted in a significant increase in GFP fluorescence and viral DNA compared with infections using fibroblast-propagated virus TB40-BAC4_fib. Infection still occurred using TB40-BAC4_fib, however. The spread of fluorescence was significantly higher in 3D tissues infected with TB40-BAC4_epi and likely includes syncytial formation as observed in the infection of monolayer NPC cultures. We initially hypothesized that this difference in infection efficiency could be explained by an increase in the cellular receptors for the virion pentamer complex. However, contrary to this hypothesis, our analyses using flow cytometry and immunofluorescence determined that NPCs and cells from cortical organoids express similar levels of pentameric receptors (i.e., Nrp2, CD46, and THBD) compared with trimeric receptors (i.e., PDGFRα and TGFβRIII). These data suggest that HCMV tropism for neural tissues is driven by the source of virus in contrast to the availability of specific entry receptors.

Following virion attachment to the entry receptors, HCMV requires glycoprotein B (gB) to promote virion fusion with the host cell membrane (12, 28, 31). Previous studies have found that neutralizing antibody against viral glycoproteins can limit the spread of the virus and might be viable therapeutic targets (29, 56). Specifically, treatment with gB antibodies was shown to block in vitro infection of both epithelial cells and fibroblasts as well as block congenital infection in a guinea pig model (44), whereas experiments using antibodies to gH had mixed neutralizing activity based on the epitopes targeted by the antibody and the cell type of interest (29). With this in mind, we tested the capacity of gB- and gH-neutralizing antibodies to limit infection in human stem cell-derived neural tissues. Previous studies from Sun et al. (27) have shown that blocking viral entry can aid organoid structure and electrophysiological function. We expanded on these findings by showing that addition of antibodies targeting viral entry at the time of infection significantly reduced viral DNA, RNA, and protein expression levels. Neutralizing antibodies partially protected NPCs resulting in near mock RNA levels of expression of neurodevelopmental transcription factors EMX1, FOXG1, DMRTA2, and FEZF2 shortly after infection. Even partial protection in transcription factor expression could have significant impacts on the developing brain (57, 58). It is important to note that we did observe a slight decrease in expression of FOXG1 and EMX1 when adding antibody in the absence of infection appearing at 48 or 72 h post treatment. It is reasonable to suggest that antibody alone is non-specifically associating with cells and influencing expression. Although more work is needed to elucidate any signaling changes tied to the antibody, this condition always maintained higher expression of any tested gene target compared with virus alone. EMX1, FOXG1, DMRTA2, and FEZF2 all play important roles in telencephalic development, progenitor cell fate decisions, cortical neuron differentiation, and overall brain structure (47, 48, 59 – 67). Previous studies from us and others have found that HCMV infection severely restricts the differentiation potential of neural progenitor cells and alters the organoid structure (17 – 19, 22, 27), and we propose that these changes and likely others are linked to downregulation of key neurodevelopmental gene networks during the complex process of neurogenesis.

Considering that the gB neutralization paradigm used here provided only transient effects, it is reasonable to suggest that complementary therapeutic approaches can be used to improve the health of neural tissues during infection. In this regard, we focused on BDNF because of its known neurotropic properties including promoting neuronal differentiation, survival, metabolism, synaptic function, and repair (51, 54, 68, 69). BDNF binds to the TrkB receptor leading to signal transduction cascades that culminate in the activation of CAMKII and MAPK signaling pathways promoting cell survival, neurogenesis, and resistance to cellular stress (70). BDNF is currently being tested as an intervention to support neuron health in neurological diseases including Alzheimer’s disease (69, 71). Moreover, previous studies have shown significant downregulation of BDNF expression with CMV infection (18, 55) which could contribute to infection-induced neuronal damage (50). The inclusion of BDNF with anti-HCMV neutralizing factors showed increased neurodevelopmental gene expression over antibody treatment alone. However, there are multiple neurotrophic factors that have similar pro-survival functions as BDNF that could also be explored in the context of HCMV infection. For example, glial-derived neurotrophic factor (GDNF) activates PI3K/AKT and MAPK signaling pathways to promote cell survival (72), and we have previously found GDNF promotes neuron survival in other models of neurodegeneration (73 – 75). Additionally, GDNF also promotes fate specification, neuron maturation, neurite outgrowth, and synaptic development (72) and is being investigated in clinical trials for Parkinson’s disease and amyotrophic lateral sclerosis (76, 77). Therefore, this multi-pronged approach provides a potential therapeutic avenue for reducing viral entry as well as boosting neuronal health, function, and survival.

In summary, we have demonstrated that epithelial-derived HCMV has increased tropism for neuronal tissues, and the addition of neutralizing antibodies in combination with neurotrophic factors maintains expression of key neurodevelopmental genes in human neural cells and tissues, which could have important implications for future therapeutic development targeting the neurological deficits associated with congenital HCMV infection.

MATERIALS AND METHODS

Cell culture and virus

Two independent control iPSC lines [4.2 and AICS-0074 (SOX2)] were grown and maintained in E8 as we have previously described (22). Line AICS-0074 was obtained from the Allen Institute for Cell Science. Following purchase, the line was thawed and expanded following our described procedure (22). Neural progenitor cells were cultured as a monolayer and differentiated using a Smadi Inhibition Kit from StemCell Technologies (#08581) requiring a minimum of three passages to become patterned NPCs. Briefly, iPSC colonies were dissociated using Versene into single cells and seeded onto Matrigel-coated 6six-well plates at a concentration of 2 × 10⁶. They were fed with NPC induction media along with a 10-μM ROCK Inhibitor (Biogems, 1293823-10MG). The following day, the media was removed, the cells were washed with PBS, and fresh NPC induction media were added. This process was repeated at every feeding, with cells being split every 6–7 days and replated at 2 × 10⁶. After three passages, the cells have become NPCs based on transcriptional, morphological, and functional analyses. The cells can then continually be passaged up to 15 times beyond the three required to form NPCs. Cerebral organoids were generated using a cortical organoid kit from Stem Cell Technologies (#08570) that relies on an established protocol (38). On day 0 of organoid culture, iPSCs were dissociated with Versene (Life Technologies, 15040066) and seeded in suspension at 90,000 cells/mL into a round bottom 96-well plate. From day 0 to day 5, cells were cultured in EB formation media with 10-μM ROCK Inhibitor added at day 0 and 100 μL fresh media added on days 2 and 4. On day 5, EBs were transferred to a low-attachment 24-well plate and fed induction media. Then, at day 7, the developing organoids were embedded in a Matrigel dome, and after 30 min of polymerization, the droplets were lifted and transferred to a low-attachment six-well plate. From there, the developing organoids were fed neural expansion/induction media, then at day 10, these media were removed, and maturation media were added. Also, at day 10, the organoids were moved onto an orbital rocker within the incubator to keep them moving as they mature. Organoids were then cultured in maturation media with media changes every 3 days.

Viral stocks were generated using the HCMV TB40/E-BAC4 genome recombined to express the EGFP gene driven by an SV40 promoter (40, 78). Isolated BAC DNA was electroporated into MRC-5 fibroblast cells as previously reported (40, 78). The resulting virus was subsequently used to make high-titered viral stocks by infecting MRC-5 fibroblasts resulting in virus TB40-BAC4_fib or by infecting ARPE-19 epithelial cells. These initial epithelial or fibroblast viral stocks are labeled as passage one (P1) epithelial or fibroblast-derived viral stocks. To improve tropism and increase titer value, the P1 viral stocks are used to infect a second set of ARPE-19 epithelial or MRC-5 fibroblast cells. These stocks are passage two (P2)-derived stocks and will have a higher viral titer compared with P1. Finally, the P2 epithelial or fibroblast-derived stocks are used to infect a third set of ARPE-19 epithelial or MRC-5 fibroblast cells making a passage 3 (P3) stock. The P3 viral stocks have increased in tropism over several passages and have higher viral titers. For stock preparations, culture supernatants were cleared of cell debris, concentrated on a sorbitol cushion, and resuspended in media. Viral stock titers were determined on the cell line that they originated from, meaning TB40-BAC4_epi titered on ARPE-19 cells and TB40-BAC4_fib titered on MRC-5 cells. Titers for viral stocks were determined in 96-well plates through a limiting dilution assay, TCID₅₀. GFP-positive wells were determined at 2 weeks post infection and resulting titers were reported in infectious units (IU) per milliliter. These concentrations were used in calculating MOI. For organoids, the MOI units used were IU/µg, while for NPCs, the units were IU/cell.

Organoids on day 30 were infected with TB40-BAC4_epi. Each organoid was weighed and infected with 500 infectious units/µg of virus for 2 h on a rocker and was generated from three separate differentiations using one WT iPSC line. Following infection, media were changed, and organoids were washed with PBS. After infection, media were replaced every 2 days. At 14 days post-infection, organoids were either fixed for cryosectioning or dissociated in Accutase enzyme and lysed for protein or DNA/RNA isolation. For NPC infections, P3 cultures and beyond were plated at 1 million cells per well and generated from four differentiations using two different healthy control iPSC lines (4.2 and NPM1). The cells were then allowed to attach for 48 h prior to infection with TB40-BAC4_epi at an MOI of 1. After a 2-h infection period, virus was removed, cells were washed with PBS, and fresh media were added. Media were replaced daily. Cell pellets were then collected at 24, 48, and 72 hpi and processed for protein isolation or DNA/RNA isolation.

HCMV gB human IgG1 [Absolute Antibody, Ab01475-10.0 Anti-HCMV gB (5A6)] and HCMV gH human IgG1 [Absolute Antibody, Ab02066-10.0 Anti-HCMV gH (HCMV16)] were used for neutralization studies. Antibodies were added 1 h prior to infection into 1 mL of media, and then, fresh antibody was added during the 2-h infection period. For studies using antibody and BDNF, antibody and BDNF were added for 1 h prior to infection and for the 2-h infection period. Following the infection, cells were pelleted and collected at 24, 48, and 72 hpi or fixed for staining. BDNF was used at a final concentration of 20 ng/mL, and antibodies were added at a final concentration of 4,000 ng/mL (high condition) and 1,000 ng/mL (low condition).

Immunofluorescence

NPCs and organoids were fixed on coverslips at various time points with or without infection in 4% PFA and then washed with PBS. For surface expression staining, NPCs were incubated with primary antibodies in 0.1% BSA with 5% normal donkey serum overnight. Meanwhile, when permeabilization was required, cells were incubated in 0.2% Triton in PBS for 30 min prior to primary antibody addition and incubation overnight. The primary antibodies used were as follows: Nrp2 (R&D Systems, AF2215-SP), PDGFRα (BD Biosciences, 556001), CD46 (R&D Systems, AF2005), THBD (AbCam, ab33513), TGFβRIII (Sigma, T1940), UL123 (Shenk Lab, Clone 1B12), SOX2 (Sigma, AB5603MI), Tuj1 (GeneTex, GTX85469), and nuclear stain Hoescht (Thermo Scientific, H3570), along with a conjugated Cellbrite Orange Cytoplasmic Membrane Dye (Biotium, 30022) in Fig. 3. The appropriate fluorescent secondaries were then used in channels AF488, AF568, and AF647 and diluted in 0.1% BSA with 5% normal donkey serum. Images were captured and produced using the Zeiss LSM980 (5×, 20×, and 40× water) at various times post-infection.

Flow cytometry

For organoids, the flow was run at day 45 of differentiation across three separate organoid differentiations with the same iPSC line and at least two biological replicates per differentiation. NPCs were analyzed after passage 3 and generated from four differentiations using the two healthy control iPSC lines. For each sample, 10,000 events were recorded, and biological replicates were averaged. Dissociation was done using Accutase (StemCell Technologies, 07920) enzyme to lift the monolayer of NPCs or dissociate the 3D structure of the organoid; the lysates were then filtered to remove debris and non-dissociated cells. After dissociation, the cells were counted and prepped for flow using the same receptor antibody targets listed above (Nrp2, PDGFRα, CD46, THBD, and TGFβRIII) along with conjugated SOX2 (BD Biosciences, 561610). Secondaries of AF488 and AF568 were used in correspondence with the correct primaries. Samples were run on the LSRII machine with 10,000 events recorded per group. The graphics and subsequent analysis were conducted using the program FlowJo.

Nucleic acid analyses

DNA samples were collected at time points indicated post-infection from NPCs or organoids using a Qiagen Kit (Qiagen, 69506). Following isolation, DNA polymerase chain reaction (PCR) was run using primer sets for UL123 and GAPDH; then, the ratio of Cq values was calculated and reported in the graphs as Viral/Host genomes. Total RNA was isolated (Qiagen, 74136) and reverse transcribed into cDNA using the Promega RT Kit (Promega, A3500). Reverse transcription quantative PCR (RT-qPCR) was performed using specific primer sequences for FOXG1 (F-CGTTCAGCTACAACGCGCTCA, R-CAGATTGTGGCGGATGGAGTT), EMX1 (F-AGCCCCGTCTTAATGCAACA, R-CTAGGATTGCGGGGCTAGTG), FEZF2 (F-GTGGTGGAATTCGCCGCCGCCATGGCCAGCTCAGCTTCCCTGGAGACCATGGTG,

R-TGCTGGATATCAGCTCTGAACTGTCCTGGCTAGGTCCTTTGCTGA), DMRTA2 (F-GCCTGCCTACGAAGTCTTTGGCTCGGTTT, R-CGTCTTGGGAAACAGATCAAACTTCTG), UL122 (F-ACCTGCCCTTCACGATTCC, R-ATGGTTTTGCAGGCTTTGATG), UL123 (F-GCCTTCCCTAAGACCACCAAT, R-ATTTTCTGGGCATAAGCCATAATC), UL44 (F-GCCCGATTTCAATATGGAGGTCAG, R-CGGCCGAATTCTCGCTTTC), and UL99 (F-GTGTCCCATTCCCGACTCG, R-TTCACAACGTCCACCCACC). The resulting Cq values were normalized to GAPDH.

Statistical analysis

Conclusions were made based on data collected from a minimum of three independent differentiation and infections with at least two technical replicates per differentiation followed by one-way ANOVA or Student’s t-test as appropriate with Tukey’s post hoc test. Data are presented as mean with SD; P < 0.05 was considered significant.