key: cord-0309266-e6gu2bpu
authors: Branscome, Heather; Khatkar, Pooja; Sharif, Sarah Al; Yin, Dezhong; Jacob, Sheela; Cowen, Maria; Kim, Yuriy; Erickson, James; Brantner, Christine A.; El-Hage, Nazira; Liotta, Lance A.; Kashanchi, Fatah
title: Retroviral Infection of Human Neurospheres and Use of Stem Cell EVs to Repair Cellular Damage
date: 2021-08-09
journal: bioRxiv
DOI: 10.1101/2020.12.31.424849
sha: e1e5eb86fe378597dcde9d44103f02ec8a89b00a
doc_id: 309266
cord_uid: e6gu2bpu

HIV-1 remains an incurable infection that is associated with substantial economic and epidemiologic impacts. HIV-associated neurocognitive disorders (HAND) are commonly linked with HIV-1 infection; despite the development of combination antiretroviral therapy (cART), HAND is still reported to affect at least 50% of HIV-1 infected individuals. It is believed that the over-amplification of inflammatory pathways, along with release of toxic viral proteins from infected cells, are primarily responsible for the neurological damage that is observed in HAND; however, the underlying mechanisms are not well-defined. Therefore, there is an unmet need to develop more physiologically relevant and reliable platforms for studying these pathologies. In recent years, neurospheres derived from induced pluripotent stem cells (iPSCs) have been utilized to model the effects of different neurotropic viruses. Here, we report the generation of neurospheres from iPSC-derived neural progenitor cells (NPCs) and we show that these cultures are permissive to retroviral (e.g. HIV-1, HTLV-1) replication. In addition, we also examine the potential effects of stem cell derived extracellular vesicles (EVs) on HIV-1 damaged cells as there is abundant literature supporting the reparative and regenerative properties of stem cell EVs in the context of various CNS pathologies. Consistent with the literature, our data suggests that stem cell EVs may modulate neuroprotective and anti-inflammatory properties in damaged cells. Collectively, this study demonstrates the feasibility of NPC-derived neurospheres for modeling HIV-1 infection and, subsequently, highlights the potential of stem cell EVs for rescuing cellular damage induced by HIV-1 infection.

6 confer beneficial effects in the context of CNS infection, including HIV-1 and HAND, remains to be addressed.

Thus, studies of this nature are imperative to better characterize the potential neuroprotective properties of stem cell 139 EVs and to determine the mechanisms by which these EVs may repair cellular damage that is induced by different 140 viral infections.

Here, we report the generation of neurospheres from NPCs derived from normal human iPSCs. These 142 cultures not only exhibit reproducibility with respect to morphology and size but also demonstrate the ability to be 143 stably maintained in culture to permit differentiation into different CNS cell types. Importantly, our data suggests 144 that these cultures are permissive to HIV-1 infection. Additionally, we include data to show that neurospheres co- 

As part of routine quality control testing, the NPCs used in these experiments were previously To assess the susceptibility of NPC-derived neurospheres to HIV-1 infection, neurospheres were cultured 208 in the presence of HIV-1 dual tropic 89.6 (MOI:10). In addition, since HIV-1 infected individuals utilize combined 209 antiretroviral therapy (cART), neurospheres were also treated with a cocktail consisting of equal parts lamivudine, 210 tenofovir disoproxil fumarate, emtricitabine, and indinavir to model the effects of cART [1] . After a period of 211 fourteen days in culture, supernatants were collected and neurospheres were harvested for downstream assays. To 212 determine whether exposure of neurospheres to HIV-1 resulted in effective viral replication, total cellular protein 213 was isolated and western blot was performed to assess the expression of different viral proteins. As shown in Fig.   214 2a, uncleaved Gag polyprotein Pr55, p24 capsid protein, and accessory protein Nef were all detected in neurospheres 215 exposed to HIV-1 89.6.

The cART cocktail used for these experiments consisted of a mixture of nucleoside reverse transcriptase 217 inhibitors (NRTIs) and a protease inhibitor; as expected, culturing of HIV-1 neurospheres with cART reduced the 218 expression of Pr55, p24, and Nef to almost undetectable levels (Fig. 2a) . RT-qPCR for HIV-1 TAR and env 219 confirmed the presence of these RNAs in HIV-1 89.6-exposed neurospheres with approximately 10 9 and 10 7 copies 220 being detected, respectively, relative to the control (uninfected) neurospheres (Fig. 2b) . Treatment of HIV-1 221 neurospheres with cART decreased the expression of both TAR and env RNA relative to the HIV-1 sample, with a 222 9 significant reduction in the expression of TAR.

To validate these results, we also exposed neurospheres to two other HIV-1 lab adapted strains, JR-CSF 224 (MOI:10) and CHO40 (MOI:10) with or without cART. After seven days in culture, supernatants were collected 225 and neurospheres were harvested for downstream assays. Total cellular proteins were isolated and, again, western 226 blot showed the expression of p24 capsid protein in neurospheres that had been exposed to each viral isolate (Fig.   227 2c). Moreover, the level of p24 expression appeared relatively consistent among 89.6, JR-CSF, and CHO40-228 exposed neurospheres. Collectively, this data shows that HIV-1 viral replication occurs in NPC-derived 229 neurospheres in a reproducible manner.

We were also interested in examining the potential tropism of HIV-1 in neurospheres. While CD4 + T cells 231 are the most abundantly HIV-1 infected cells in the body there is ample evidence to suggest that microglia and 232 astrocytes are not only capable of being infected by HIV-1, but may also serve as viral reservoirs, as reported 233 elsewhere [56, 57, 58, 59, 60, 61, 62] . For this experiment, NPCs were gently disaggregated with trypsin and 234 immunoprecipitation (IP) was performed using antibodies for either astrocytes (GFAP + ), microglia-like cells 235 (CD11b + ), mature neurons (GAD65 + ), or NSCs (SOX2 + ). We next performed direct PCR of the IP material without 236 purification of DNA, as RNA isolation from a relatively small number of infected cells would not yield a sufficient 237 amount of sample material for RT-qPCR. Results from qPCR confirmed the presence of TAR DNA in microglia-238 like cells in the 89.6, JR-CSF, and CHO40 infected neurospheres (Fig. 2d) . While the highest numbers of TAR 239 DNA were associated with astrocytes (~480 copies) and microglia-like cells (~780 copies) in the JR-CSF samples, 240 lower amounts were also detected in microglia-like cells from the 89.6 (~ 114 copies) and CHO40 (~ 125 copies) 241 samples. In addition, astrocytes from the JR-CSF neurospheres (~480 copies) had substantially higher levels of 242 TAR DNA than astrocytes from the 89.6 and CHO40-infected neurospheres (< 20 copies each 

Based on this data, and given that microglia are the prominent innate immune cells in the CNS and, 246 therefore, represent a major target of HIV-1, we next attempted to increase the population of myeloid cells in 247 neurospheres to better assess the replication potential of HIV-1. To this end, differentiated neurospheres were 248 treated with the following cytokines: IL-34 (100 ng/mL), M-CSF (25 ng/mL), TGFβ-1 (50 ng/mL) every other day 249 for a period of seven days, as previous studies have demonstrated the ability of these cytokines to induce microglia 250 10 differentiation in iPSCs [16, 63, 64] . Neurospheres were then infected with HIV-1 dual tropic 89.6 (MOI:10) with 251 or without cART as described above. After fourteen days, supernatants were collected and neurospheres were 252 harvested for downstream assays. As shown in Fig. 2e , western blot revealed robust expression of p24 capsid 253 protein with undetectable expression of Pr55, thus suggesting an efficient cleavage of the viral polyprotein in the 254 induced neurospheres relative to the uninfected sample. We also observed increased expression of the intracellular 255 forms (e.g. 30, 40 kDa) of the accessory protein Nef [65] relative to the uninfected sample. Again, treatment of 256 HIV-1 neurospheres with cART decreased the expression of both p24 and Nef.

Cytokine-treated neurospheres were also assessed via western blot to evaluate the expression of different 258 microglia markers (e.g. IBA-1, CD11b, CD45). Data in Fig. 2e indicates that IBA-1 expression was relatively 259 consistent among each sample, whereas slightly varying levels of expression of CD11b and CD45 were observed 260 between samples. Additionally, the expression levels of apoptotic proteins were evaluated to assess whether 261 cytokine treatment or cART had any potential effects on apoptosis (Fig. 2e) . Here, expression of the pro-apoptotic 

RT-qPCR for HIV-1 TAR and env confirmed the expression of these RNAs in the HIV-1 cytokine-treated 267 samples. Of note, similar amounts of both TAR and env (~ 10 8 copies each) were detected, suggesting that 268 increased viral processivity was occurring in these cultures (Fig. 2f) . Again, treatment with cART resulted in 269 significant reductions of both TAR and env RNA relative to the HIV-1 sample. Collectively, this data suggests that 270 treatment of neurospheres with IL-34, M-CSF, and TGFβ-1 effectively enhanced viral replication and processivity 271 and that this may be attributed to the induction of a pro-myeloid phenotype. This data also indicates that cytokine 272 treatment alone, as well as cART alone, did not promote apoptosis in neurospheres.

We next sought to assess the potential effects of HIV-1 (89.6) on the expression levels of multiple proteins 274 or enzymes associated with different CNS cell types. To this end, western blot was performed to evaluate the 275 relative expression microglia-like cells (CD11b + , CD163 + ), astrocytes (GFAP + , glutamine synthetase + ), 276 dopaminergic neurons (TH + , FOXA2 + ), GABAergic neurons (GAD65 + , GAD67 + ), glutamatergic neurons (BNPI + )

and NSCs (SOX2 + ). As shown in Fig. 3a , HIV-1 appeared to increase the expression of differentiated neurons, 278 particularly TH + , GAD65 + /GAD67 + , and BNPI + neurons, whereas treatment with cART restored the expression of 279 these markers to levels that were comparable to the uninfected sample. A similar pattern of expression was also 280 observed for GFAP, while the expression of CD11b appeared relatively consistent between the uninfected, HIV-1 281 infected, and cART samples. On the other hand, expression of CD163 appeared to slightly decrease in both the 282 HIV-1 and cART samples relative to the uninfected sample.

As an additional control, we also evaluated the effects of co-culturing neurospheres with the HTLV-1 

To confirm whether HTLV-1 was capable of replicating in neurospheres, western blot was performed to 298 assess the expression of HTLV-1 Tax and matrix (p19) proteins. As shown in Fig. 3b , the expression of both Tax 299 and p19 was higher after co-culturing of neurospheres with irradiated HUT 102 cells (relative to the background 300 control), thereby indicative of increased viral replication. As expected, treatment with IFN/AZT reduced the 301 expression of Tax and p19 to levels that were similar to the background control. Furthermore, RT-qPCR for HTLV-302 1 tax, env, and hbz RNA showed that the expression of these RNAs significantly increased after co-culture and that 303 treatment with IFN/AZT resulted in a significant reduction of both tax and env transcripts (Fig. 3c ).

Overall, these results demonstrate the ability of HIV-1 to replicate in NPC-derived neurospheres, therefore 305 indicating that these models may be permissive to infection. Moreover, our data suggests that the exposure of 

The EVs used for these experiments were isolated using a scalable platform which incorporated the use of we also include data from A549 (cancer) EVs which served as both a control for our EV isolation method as well as 327 a non-stem cell control. NTA analysis shows that the size distribution of each EV preparation falls within the range 328 of 50 to 250 nm (Fig. 4a) . Consistent with our previous research, iPSC EVs and A549 EVs appear more 329 heterogeneous in their size distribution relative to MSC EVs [76] . TEM of each EV preparation also confirmed the 330 presence of cup-shaped vesicles that appeared less than 200 nm in diameter. Representative TEM images from each 331 EV preparation are included in Fig. 4b . Interestingly, these images also showed that iPSC EVs tended to cluster 332 together in aggregates of smaller (~50 nm) vesicles whereas MSC EVs appeared as both larger (~100 nm) vesicles 333 and aggregates of smaller (~50 nm) vesicles.

To assess the efficiency of their uptake by differentiated neurospheres, EVs were fluorescently labeled with 335 a commercially available dye which stains both neutral and nonpolar lipids (BODIPY). Labeled EVs were then 336 added to cultures of differentiated neurospheres. As shown in Fig. 4c , fluorescent microscopy initially revealed a 337 relatively weak and diffuse signal that persisted throughout the first 48 hours (day 2). However; over time, the 338 fluorescent intensity increased. Representative photographs (day 8) show a strong fluorescent signal that appeared 339 to be predominantly localized in the center of each EV-treated neurosphere. We also observed similar results when 340 treating neurospheres with EVs labeled with the SYTO RNASelect nucleic acid stain ( Supplementary Fig. 2a ).

Additionally, image analysis of EV-treated, cross-sectioned neurospheres pointed towards a strong and relatively 

EVs [76] . Again, the purpose of including A549 EVs was to serve as a non-stem cell control. Here, we share an 393 expanded set of data from our initial studies to highlight the differences that exist among these different EV 394 populations.

Overall, approximately 6.5 x10 7 , 1.9 x 10 7 , and 2.6 x 10 7 valid reads were obtained from iPSC, MSC, and 396 A549 EVs, respectively, with each preparation having a Q30 score greater than 85% (Supplementary Table 1 ). Of 397 these reads, approximately 80.88% (iPSC EVs), 71.21% (MSC EVs), and 70.66 (A549 EVs) were mapped to the 398 genome. As shown in Fig. 6a , the majority of lncRNAs among each EV preparation mapped to introns. iPSC EVs 399 had the highest number of intronic transcripts (93%) followed by MSC EVs (79%) and A549 EVs (75%). The 400 percentage of transcripts mapping to either exons or intergenic regions ranged from approximately 10 to 15% among 401 MSC and A549 EVs; however only ~3% of transcripts from iPSC EVs were mapped to these regions.

Transcripts that were not annotated in genome annotation databases were subsequently designated as novel Data in Fig. 6b shows that within each EV prep, the majority of transcripts either fell within a reference intron (i) 407 (> 50%) or were classified as unknown intergenic transcripts (u) (< 40%) and, moreover, that ≤ 0.1% of transcripts 408 among each sample were classified as either "j", "o", or "x". and that their expression was significantly increased in iPSC EVs relative to MSC EVs (Fig. 7b) .

To assess whether the identified sequences were capable of binding to PKR the biotin-conjugated synthetic 447 RNA sequences were incubated with whole cell lysate from the neuronal cell line SHSY5Y and then purified using 448 Streptavidin-Sepharose beads. Western blot was performed to assess the expression of PKR, as well as other critical 449 RNA binding proteins. The results in Fig. 7c confirm that the relevant sequences from both ADIRF-AS1 and 450 AC120498.9 were capable of binding not only to PKR, but also to other double stranded RNA (dsRNA) binding 451 proteins including DICER and ADAR1. We additionally performed similar experiments using whole cell lysate 452 from either astrocytes or MDMs ( Supplementary Fig. 6 ). Again, western blot analysis confirmed expression of 453 PKR, DICER, and RIG-I, thereby highlighting the potential of these RNAs to bind to multiple different RNA 454 binding proteins in CNS relevant cell types. As a negative control for these experiments, we did not detect 455 expression of p65 (data not shown). Collectively, we suspect that these innate immune molecules may bind to EV-456 associated lncRNAs to either activate their enzymatic activities or, alternatively, to suppress their functions in 457 recipient cells. Future studies will better focus on the potential significance of this binding. 

For this assay SHSY5Y cells were treated with EVs from either U1 cART-treated cells or a combination of 468 U1 cART EVs with stem cell EVs. Representative images in Fig. 7d show that after approximately eight days in 469 culture the untreated cells appeared healthy, well-attached, and exhibited a spindle-like morphology. Conversely, 470 neurons exposed to U1 cART EVs displayed an altered appearance that was characterized by a reduction in 471 attachment and clumping/ aggregation of the remaining attached cells. However, treatment with either iPSC or 472 MSC EVs at an approximate ratio of 1:1000 (recipient cell: EV) promoted the re-attachment and spreading of cells 473 that more closely resembled the untreated control culture. To ascertain whether these effects may be mediated by 474 18 EV-associated RNA, stem cell EVs were exposed to UV(C) irradiation to deactivate RNA. Treatment of neurons 475 with irradiated EVs appeared to abrogate the morphological effects associated with non-irradiated stem cell EVs.

Cellular viability was next assessed using a luminescence-based assay. As shown in Fig. 7e , treatment of neurons 477 with U1 cART EVs resulted in a slight decrease in viability relative to the control. While the treatment with stem 478 cell EVs did improve cellular viability, only iPSC EVs were associated with a significant increase. Furthermore, 479 there were no significant effects of UV(C) irradiated EVs on cellular viability.

In a separate experiment, we have also utilized EVs isolated from cells infected with the human 481 coronavirus OC43 to induce neuronal damage ( Supplementary Fig. 7) . Similar to what was observed with U1 cART 482 EVs, neurons exposed to OC43 EVs detached from the substrate and appeared round in morphology. Again, 

[111]. Importantly, we also have data demonstrating that our NPC-derived neurospheres, which were generated 513 from human iPSCs, express both CD4 and CCR5. Moreover, this expression remained relatively consistent after 514 treatment of infected samples with cART and EVs; this data further supports our claims regarding the susceptibility 515 of neurospheres to HIV-1 infection (Supplementary Fig. 8 ). To the best of our knowledge, this is the first study 516 utilizing human NPC-derived neurospheres for retroviral infection.

The use of well-characterized and qualified material, as well as the establishment of robust and 518 reproducible protocols is essential for ensuring the integrity of scientific data. The protocols employed here 519 permitted the efficient formation of neurospheres, resulting in well-defined cultures that displayed uniformity in size 520 and appearance throughout the duration of the study; thereby overcoming some of the challenges that have 521 historically been associated with use of 3D neuronal cultures [52] . The NPCs used for these experiments also 522 underwent extensive characterization to authenticate not only their identity, but also their differentiation potential.

Our data suggests that NPC-derived neurospheres maintain their ability to differentiate into both mature neurons 524 (e.g. dopaminergic, GABAergic, glutamatergic) and glial cells (i.e. astrocytes and potentially microglia) in 3D 525 cultures. It is worth noting that properly distinguishing between microglia and macrophages has historically been 526 20 challenging as these cell types are often associated with many of the same surface markers. Due to the difficulties 527 associated with their phenotypic identification, here we use the term "microglia-like cells" to correlate with positive 528 expression of either CD163, CD11b, CD45, and IBA-1 as these markers are associated with either resting or 529 activated microglia and macrophages [112, 113, 114, 115, 116] . While the potential of neurospheres to differentiate 

uncleaved Gag polyprotein Pr55, and accessory protein Nef, all of which were detected after fourteen days in 553 culture. As a measure of validation, we also detected p24 in neurospheres exposed to two other lab-adapted strains 554 21 (JR-CSF and CHO40). IP assays further revealed that the highest copies of HIV-1 TAR DNA were associated with 555 microglia-like cells. The JR-CSF infected sample also displayed relatively high copy numbers of TAR DNA 556 associated with astrocytes; a finding that is not entirely surprising given that this viral isolate originated from the 557 cerebrospinal fluid of an infected patient.

Perhaps more importantly, our data suggests that treatment with IL-34, M-CSF, and TGFβ-1 for a period of 559 one week had the potential to induce a microglia-like phenotype in neurospheres. Notably, we also found that can, however, be addressed in "3D" neurosphere models, especially if we are able to maintain long-term cultures 580 over a period of three to six months prior to genome analysis.

Interestingly, we observed unique expression profiles of cell-specific markers among uninfected and 582 22 infected neurospheres (Fig. 3a) . First, these results are in general agreement with our previous assays (ICC and IP 

Exposure to IR is known to initiate cell cycle arrest; moreover, we have recently published data using IR to activate 

The results from our uptake assays first confirmed that EVs were capable of penetrating neurospheres. suggests that A549 (cancer) EVs may also exert anti-apoptotic and anti-inflammatory properties, further 650 interpretation and analysis of these results is outside the scope of the current manuscript. Unlike stem cell EVs, the 651 use of cancer EVs for potential repair is not well-studied and the long-term functional outcomes associated with 652 exposure to cancer EVs needs to be better characterized.

Although the results from these experiments aren't cell-type specific, our future experiments will more 654 thoroughly examine which cell types in neurospheres are associated with increased levels of apoptosis.

Additionally, our future experiments will focus on examining the expression levels of other critical mediators of 656 apoptosis, including p53 and its kinases (e.g. ATM, ATR), as well as a broader range of proteins involved in the 657 early, middle, and late stages of apoptosis to better define the reparative mechanisms of stem cell EVs.

EVs are associated with a rich diversity of cargo which includes various non-coding RNAs (i.e. small non- 

Furthermore, the actual copy number of these RNAs in various subpopulations of EVs (e.g. those ranging from ~30 669 to 220 nm in size) requires further purification, followed by functional repair assays and then proteomics and RNA 670 seq. This will better define which EV subtype contains cytokines, proteins, and RNAs that map to a certain 671 chromosome. Therefore, future experiments will better define the type of EVs needed to repair, as well as their 672 protein and RNA cargo that may functionally regulate either the host genome, splicing in the nucleus, translation 673 inhibition, or mRNA degradation in the cytoplasm.

The identification of at least two iPSC EV-associated lncRNAs (i.e. AC120498.9, ADIRF-AS1) whose 

Therefore, we speculate that these structures could potentially serve as RNA sponges for cellular miRNA that 686 regulate pathways such as apoptosis, cell cycle, and cytokine secretion. Future experiments will better define how 687 these stem and loop RNA structures are capable of binding to dsRNA proteins and how they may also regulate 688 miRNAs that may be essential for cellular repair in damaged cells.

It has been shown that EVs released from HIV-1 infected cells contain viral by products including Nef, Tat immune response to undergo cell cycle arrest. In this scenario, EV-associated lncRNAs could prolong this effect by 716 binding to, and thus inactivating, the host innate immune molecules to prevent further viral replication. In turn, this 717 would also dampen the production of damaging EVs from infected cells. Over time, the EV-associated cytokines 718 and growth factors will promote cellular replication of both infected and uninfected cells. However, we believe that 719 the net outcome will confer an overall protective effect, since we suspect that uninfected cells represent a majority 

To further induce differentiation in neurospheres, the fully supplemented differentiation medium (listed 766 above) was spike with the following cytokines: IL-34 (100 ng/mL), M-CSF (25 ng/mL), and TGFβ-1 (50 ng/mL).

Neurospheres were cultured in this medium for an additional seven days, receiving fresh treatments every two to 768 three days, prior to HIV-1 infection. 

U1 cART EVs were isolated via ultracentrifugation. First, supernatants were centrifuged at 1,200 rpm for 909 five minutes to pellet any cellular debris. The resulting supernatant was then centrifuged at 100,000 x g for 90 910 minutes using the 70Ti rotor (Beckman). The pellet was washed with PBS followed by another centrifugation at 911 100,000 × g for 90 minutes. EV pellets were resuspended in PBS and then frozen at -20°C for downstream analysis.

A549 cells were infected with 5.0 x 10 5 TCID50 of human coronavirus OC43 for 96 hours. Supernatants 913 were collected and centrifuged at 2,000 rpm for five minutes, filtered through a 0.22 µm sterile filter, and then 914 centrifuged at 100,000 x g for 90 minutes using the 70Ti rotor (Beckman). The pellet was resuspended in sterile 915 PBS and treated with 45 mJ/cm 3 of UV(C) LED irradiation. Samples of UV(C)-treated EVs (1, 5, 10 µL) were 916 added to Vero cells for two passages and no viral outgrowth was detected (data not shown). Remaining EVs were 917 frozen at -20°C for downstream analysis. U1 cART and OC43 EVs were analyzed via NTA using the ZetaView Z-NTA (Particle Metrix). Prior to 919 analysis the equipment was calibrated per the manufacturer's recommendation and each EV sample was diluted in 920 deionized water prior to analysis. The acquired data was analyzed by the instrument's built-in software.

Transmission Electron Microscopy (negative staining method)

To prepare EV samples for electron microscopy, different dilutions of each sample were prepared and 5 µL 923 was adsorbed to the formvar side of a glow-discharged, 200 mesh copper grid with formvar and carbon coating.

Excess sample was removed from the grids with filter paper. Samples were fixed with 4% glutaraldehyde in 0.12 M 925 sodium cacodylate buffer (pH 7.2) for five minutes, washed four times with distilled, deionized water, and stained 926 with 1% aqueous uranyl acetate for 60 seconds. The uranyl acetate was removed with filter paper and allowed to air 927 dry. Imaging was performed using an FEI Talos F200X transmission electron microscope at 80KV (Thermo Fisher 

Total RNA was extracted from EVs using TRIzol reagent (Invitrogen) following the manufacturer's 939 recommendations. RNA sequencing and sequence analysis was performed by LC Sciences, LLC. Briefly, the 2100 940 Bioanalyzer System (Agilent) was used to analyze total RNA and ribosomal RNA was removed per the Ribo-Zero 941 Gold Kit (Illumina). The remaining RNA fragments were used to construct a strand-specific cDNA library using the 942 dUTP method (average library insert size was 300 ± 50 bp) and pair-end sequencing was performed using the Hiseq 943 4000 (Illumina) following the manufacturer's recommended protocol. Reads containing adaptor contamination and 944 low quality bases were removed with Cutadapt and in-house perl scripts [154] . FastQC was used to verify transcript 945 35 quality, Tophat2 [155] and Bowtie 2 [156] were used for mapping reads to the genome, and in-house Perl scripts 946 were used for GO and KEGG enrichment analysis. StringTie [157] was used for transcript assembly and Ballgown

[158] was used to perform differential expression analysis. Significant terms for GO analysis were determined by a 948 hypergeometric equation and GO terms with a p value <0.05 were considered significant [159] . All data generated 949 from the RNAseq analysis represents data from one biological replicate each of A549, MSC, and iPSC EVs.

The secondary structures of selected long non-coding RNAs were predicted using the RNAfold Server of 

EVs were added to a 30% slurry of NT80/82 Nanotrap particles (Ceres Biosciences) and PBS was added to 988 bring the final volume to 500 µL. Samples were rotated at 4°C for two hours, spun, and washed with PBS.

Laemmli buffer was added to each sample, heated for ten minutes at 37°C, and then spun to pellet the Nanotraps.

The resulting supernatants were loaded on a Novex 10% Zymogram Plus (Gelatin) Protein Gel (ThermoFisher 991 Scientific). The gel was run at 120V for approximately 90 minutes, renatured for 30 minutes at room temperature 992 on a rocker, and incubated at 37°C in developing buffer for a period of 24 to 36 hours. After development, the gel 993 was gently stained for an additional 24 to 48 hours with Coomassie blue R-250. Imaging was performed with the 994 ChemiDoc Touch Imaging System (Bio-Rad). Branscome et al. Fig. 1a-c 

Week 1-2 3-5 6-8 9-11 12-14

Week 12

Week 13

Week 14

UNAIDS. Global HIV & AIDS statistics -2019 fact sheet. PDF at 1026

HIV-associated Neurocognitive Disorder--Pathogenesis and Prospects for Treatment

Neuropathogenesis of HIV-associated Neurocognitive Disorders: Roles for 1030 Immune Activation, HIV Blipping and Viral Tropism

HIV-1-associated Neurocognitive Disorder: Epidemiology, Pathogenesis, Diagnosis, and 1032 Treatment

The Impact of HIV-1 on Neurogenesis: Implications for HAND

Viral and Cellular Factors Underlying Neuropathogenesis in HIV Associated 1036 Neurocognitive Disorders (HAND)

HIV-associated Synaptic Degeneration

Human Immunodeficiency Virus Type 1 gp120 and Tat Induce 1039

Mitochondrial Fragmentation and Incomplete Mitophagy in Human Neurons

Personalized Nanomedicine for CNS Diseases. Drug Discov 1041 Today

A Systematic Analysis for the Global Burden of Disease Study

Induced Pluripotent Stem Cell Technology: A Decade of Progress

Induction of Pluripotent Stem Cells From Mouse Embryonic and Adult 1047

Induction of Pluripotent Stem Cells From Adult Human Fibroblasts by Defined Factors

Differentiation of Human Oligodendrocytes From Pluripotent Stem Cells

Clinical Trials in a Dish: The Potential of Pluripotent Stem Cells to Develop 1053 Therapies for Neurodegenerative Diseases

iPSC-Derived Human Microglia-like Cells to Study Neurological Diseases

Preparation and Co-Culture of iPSC-Derived Dopaminergic Neurons and Astrocytes

Replication of the Zika Virus in Different iPSC-derived Neuronal Cells and Implications to 1059

Modeling Sporadic ALS in iPSC-derived Motor Neurons Identifies a Potential Therapeutic 1061

Pluripotent Stem Cells for Modelling and Cell Therapy of Parkinson's 1063

Mitochondrial Dysregulation and Impaired Autophagy in iPSC-Derived 1065

Dopaminergic Neurons of Multiple System Atrophy

Inflammatory Responses to a Pathogenic West Nile Virus Strain

Alzheimer's Disease With iPSC-derived Brain Cells. Mol 1069 Psychiatry

Recent Overview of the Use of iPSCs Huntington's Disease 1071

Genetically-edited Induced Pluripotent Stem Cells Derived From HIV-1-infected Patients on 1073

Therapy Can Give Rise to Immune Cells Resistant to HIV-1 Infection

The Third Dimension Bridges the Gap Between Cell Culture and 1075 Live Tissue

A Guide to Generating and Using hiPSC Derived NPCs for the

Neurosphere Based Differentiation of Human iPSC Improves Astrocyte Differentiation. Stem 1079 Cells Int

Comparative Performance Analysis of Human iPSC-derived and Primary Neural 1081 Progenitor Cells (NPC) Grown as Neurospheres in Vitro

Regional Specification of Rodent and Human Neurospheres

Regional Specification of Neurosphere Cultures Derived 1085

From Subregions of the Embryonic Telencephalon

Strengths and limitations of the neurosphere culture system

Zika virus impairs growth in human neurospheres and brain organoids

Zika virus disrupts molecular fingerprinting of human neurospheres

Generation of three-dimensional human neuronal cultures: application to modeling CNS viral 1093 infections

Modeling HIV-1 neuropathogenesis using 1095 three-dimensional human brain organoids (hBORGs) with HIV-1 infected microglia

The need for new approaches in CNS drug discovery: Why drugs have 1097 failed, and what can be done to improve outcomes

Challenges and Opportunities in Central Nervous System Drug 1099

Review: MSC-Derived Exosomes for Cell-Free Therapy

Shedding light on the cell biology of extracellular vesicles

Extracellular Vesicles: Composition, Biological 1105 Relevance, and Methods of Study

Protein, and MicroRNA Composition Within Mesenchymal Stem Cell-Derived 1107

Potential Therapies by Stem Cell-Derived Exosomes in 1109 CNS Diseases: Focusing on the Neurogenic Niche

Induced pluripotent stem cell-derived extracellular vesicles: A novel approach for cell-free 1111 regenerative medicine

Extracellular Vesicles in Physiology, Pathology, and Therapy of the Immune and Central 1113

Nervous System, with Focus on Extracellular Vesicles Derived from Mesenchymal Stem Cells as Therapeutic 1114

Mesenchymal stem cell-derived extracellular vesicles: novel 1116 frontiers in regenerative medicine

Use of Stem Cell Extracellular Vesicles as a "Holistic" Approach to CNS Repair. Front 1118

Therapeutic Potential of Extracellular 1120

Vesicles for the Treatment of Nerve Disorders

Blocking Zika virus vertical transmission

Blood cell-derived induced pluripotent stem cells free of reprogramming factors generated by 1123 Sendai viral vectors

Large-scale generation of human iPSC-derived neural stem cells/early neural progenitor cells 1125 and their neuronal differentiation

Comparison of 2D and 3D neural induction methods for the generation of neural 1127 progenitor cells from human induced pluripotent stem cells

Characteristic analyses of a neural differentiation model from iPSC-derived neuron according to 1129 morphology, physiology, and global gene expression pattern

Rapid and efficient differentiation of functional motor neurons from human iPSC for neural 1131 injury modelling

Recapitulation of Human Neural Microenvironment Signatures in iPSC-Derived NPC 3D

HIV-1 infection of the developing nervous system: central role 1135 of astrocytes in pathogenesis

Is the central nervous system a reservoir of HIV-1?

Astrocytes as an HIV Reservoir: Mechanism of HIV Infection

HIV-1 target cells in the CNS

How HIV-1 Attacks and Resides in the Central Nervous System

Microglial Cells: The Main HIV-1 Reservoir in the Brain

HIV infects astrocytes in vivo and egresses from the brain to the periphery

A Highly Efficient Human Pluripotent Stem Cell Microglia Model Displays a Neuronal-1149

Co-culture-Specific Expression Profile and Inflammatory Response

Development and 1151 validation of a simplified method to generate human microglia from pluripotent stem cells

The use of Nanotrap particles technology in capturing HIV-1 virions and viral proteins from 1154 infected cells

HTLV-I associated myelopathy/tropical spastic paraparesis (HAM/TSP): a chronic 1156 progressive neurologic disease associated with immunologically mediated damage to the central nervous system

Human T cell leukemia virus type I and neurologic disease: events in bone marrow, peripheral 1159 blood, and central nervous system during normal immune surveillance and neuroinflammation

Role of resident CNS cell populations in HTLV-1-1162 associated neuroinflammatory disease

Neuroimmunology of Human T-Lymphotropic Virus Type 1-Associated 1164

Therapeutic doses of irradiation activate viral transcription and induce apoptosis in HIV-1 1166 infected cells

Potential of Radiation-Induced Cellular Stress for Reactivation of Latent HIV-1 1168 and Killing of Infected Cells

HTLV-1 Extracellular Vesicles Promote Cell-to-Cell Contact

Treatment of adult T-cell leukemia-lymphoma with a combination of interferon alfa and 1172 zidovudine

Brief report: treatment of adult T-cell leukemia-lymphoma with zidovudine and interferon 1174 alfa

How to Control HTLV-1-Associated Diseases: Preventing de Novo Cellular Infection Using 1176

Stem Cell Extracellular Vesicles and their Potential to Contribute to the Repair of 42

Tangential Flow Filtration for Highly Efficient Concentration of Extracellular Vesicles from 1182

Large Volumes of Fluid

Exosomes Produced from 3D Cultures of MSCs by Tangential Flow Filtration Show 1184

Higher Yield and Improved Activity

Freeze-Dried Extracellular Vesicles From Adipose-Derived Stem Cells Prevent 1186

Matrix metalloproteinase-9 facilitates 1188 remyelination in part by processing the inhibitory NG2 proteoglycan

A delicate balance: role of MMP-9 in brain development and 1190 pathophysiology of neurodevelopmental disorders

Assessment of Gelatinases (MMP-2 and MMP-9 by Gelatin Zymography

Detection of Matrix Metalloproteinases by Zymography

Adipose-derived mesenchymal stem cell-derived exosomes markedly protected the brain 1196 against sepsis syndrome induced injury in rat

Neuroprotective effect of placenta-derived mesenchymal stromal cells: role of exosomes

Exosomes Derived from Bone Mesenchymal Stem Cells Repair Traumatic Spinal Cord Injury by 1200

Suppressing the Activation of A1 Neurotoxic Reactive Astrocytes

Bone marrow mesenchymal stem cells-derived exosomes improve injury of hippocampal 1202 neurons in rats with depression by upregulating microRNA-26a expression

Exosomes derived from mesenchymal stem cells repair a Parkinson's disease model by 1205 inducing autophagy

Effect of exosomes from adipose-derived stem cells on the apoptosis of Schwann cells in 1207 peripheral nerve injury

Intranasal MSC-derived A1-exosomes ease inflammation, and prevent abnormal neurogenesis 1209 and memory dysfunction after status epilepticus

Exosomes Isolated From Human Umbilical Cord Mesenchymal Stem Cells Alleviate 1211

Neuroinflammation and Reduce Amyloid-Beta Deposition by Modulating Microglial Activation in Alzheimer's 1212

Exosomes from adipose-derived stem cells alleviate neural injury caused by 1214 microglia activation via suppressing NF-kB and MAPK pathway

Exosomes Derived From Bone Mesenchymal Stem Cells Ameliorate Early Inflammatory 1216 Responses Following Traumatic Brain Injury

hucMSC derived exosomes promote functional recovery in spinal cord injury mice via 1218 attenuating inflammation

Exosomes derived from umbilical 1220 cord mesenchymal stem cells reduce microglia-mediated neuroinflammation in perinatal brain injury

Schwann cell-like differentiated adipose stem cells promote neurite 1223 outgrowth via secreted exosomes and RNA transfer

Mesenchymal stem cell-derived extracellular vesicles affect disease outcomes via transfer of 1225 microRNAs

Extracellular vesicles derived from human Wharton's jelly mesenchymal stem cells 1227 protect hippocampal neurons from oxidative stress and synapse damage induced by amyloid-β oligomers

Exosomes from mesenchymal stem/stromal cells: a new therapeutic paradigm

Mesenchymal Stem Cells Reduce Corneal Fibrosis and Inflammation via Extracellular 1232

Vesicle-Mediated Delivery of miRNA

Human bone marrow-and adipose-mesenchymal stem cells secrete exosomes enriched in 1234 distinctive miRNA and tRNA species

The microRNA regulatory landscape of MSC-derived exosomes: a systems view

Small non-coding RNA landscape of extracellular vesicles from human stem cells

Activation of PKR: an open and shut case?

Analysis of high-affinity binding of protein kinase R to double-stranded 1241 RNA

Exosomes derived from HIV-1-infected cells contain trans-activation response element 1243 RNA

Exosomes from HIV-1-infected Cells Stimulate Production of Pro-inflammatory 1245

Cytokines through Trans-activating Response (TAR) RNA

Exosomes from uninfected cells activate transcription of latent HIV-1

Comparative analysis of human 1251 microglial models for studies of HIV replication and pathogenesis

CD163 identifies a unique population of ramified microglia in HIV 1253 encephalitis (HIVE)

Markers of microglia in post-1257 mortem brain samples from patients with Alzheimer's disease: a systematic review

Pro-inflammatory activation of microglia in the brain of patients with sepsis

Differential transcriptional profiles identify microglial-and macrophage-specific 1262 gene markers expressed during virus-induced neuroinflammation

Microglia, macrophages, perivascular macrophages, and pericytes: a review of 1264 function and identification

HIV-1 infection and AIDS: consequences 1266 for the central nervous system

AIDS dementia is associated with massive, activated HIV-1 infection and 1268 concomitant expression of several cytokines

Cellular reservoirs of HIV-1 in the central nervous system of infected individuals: 1270 identification by the combination of in situ polymerase chain reaction and immunohistochemistry

Localization of HIV-1 in human brain using polymerase chain reaction/in situ 1273 hybridization and immunocytochemistry

Astrocytic expression of HIV-1 Nef impairs spatial and recognition memory

A sensitive, quantitative assay for 1277 human immunodeficiency virus type 1 integration

A more precise HIV integration assay designed to detect small differences finds lower levels of 1279 integrated DNA in HAART treated patients

Detecting HIV-1 integration by repetitive-sampling Alu-gag PCR

Measuring integrated HIV DNA ex vivo and in vitro provides insights about 1283 how reservoirs are formed and maintained

Glial fibrillary acidic protein: GFAP-thirty-one years

Astrocytes: biology and pathology

Astrocyte activation and dysfunction and neuron death by 1288 HIV-1 Tat expression in astrocytes

Impact of HIV on regional & cellular organisation of the brain. Curr 1290 HIV Res

Involvement of p300 in constitutive and HIV-1 Tat-activated expression of glial fibrillary 1292 acidic protein in astrocytes

Régulation du cycle cellulaire des cellules exposées aux 1294 radiations ionisantes

Antiretroviral Drugs Alter the Content of Extracellular Vesicles from HIV-1-Infected 1296

Ebola Virus VP40 Modulates Cell Cycle and Biogenesis of Extracellular Vesicles

Purification of High Yield Extracellular Vesicle Preparations Away from Virus

The Therapeutic Potential of Mesenchymal Stem Cell-Derived Exosomes in Treatment of 1302

Specific proteolytic cleavage of 1304 poly(ADP-ribose) polymerase: an early marker of chemotherapy-induced apoptosis

ADP-ribose) polymerase-1 cleavage during apoptosis: an update

BCL-2 family proteins: changing partners in the dance towards death

Caspases: the executioners of apoptosis

Emerging roles of caspase-3 in apoptosis

Exosome and exosomal microRNA: trafficking, sorting, and function

Extracellular vesicle associated long non-1315 coding RNAs functionally enhance cell viability

Exosomes as miRNA Carriers: Formation-Function-Future

Exosomal long non-coding RNAs as 1319 biomarkers in human diseases

The Long and Short of It: The Emerging Roles of Non-Coding RNA in Small 1321

Sorting Mechanisms for MicroRNAs into Extracellular Vesicles and Their Associated

Circular RNAs: A Novel Class of Functional RNA Molecules with a Therapeutic 1325 Perspective

Circular RNAs function as ceRNAs to regulate and control human cancer progression. Mol 1327 Cancer

Genetic characterization of HIV type 1 Nef-induced vesicle secretion

The Role of Extracellular Vesicles in Viral Infection and Transmission. Vaccines (Basel)

Extracellular vesicles from HTLV-1 infected cells modulate target cells and viral spread

Analysis of relative gene expression data using real-time quantitative PCR 1335 and the 2(-Delta Delta C(T)) Method

Cutadapt removes adapter sequences from high-throughput sequencing reads

TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene 1339 fusions

Fast gapped-read alignment with Bowtie 2

StringTie enables improved reconstruction of a transcriptome from RNA-seq reads

Ballgown bridges the gap between transcriptome assembly and expression analysis

edgeR: a Bioconductor package for differential expression 1346 analysis of digital gene expression data

Hofacker, I.L. The Vienna RNA websuite

Optimal computer folding of large RNA sequences using thermodynamics and 1350 auxiliary information

Expanded sequence dependence of thermodynamic 1352 parameters improves prediction of RNA secondary structure

analysis of variance (ANOVA) using multiple comparisons. To justify the use of ANOVA, experimental data was 998 37 assumed to approximate a normal distribution. P values were defined as follows: of greatest statistical significance 999 (< 0.0001), of greater significance (< 0.005), and significant (< 0.05).

We would like to thank all members of the Kashanchi lab for their contributions, especially Gwen Cox and 1002 interns Gabrielle Heller and Sahan Raghavan. We also would like express thanks to ATCC management, 1003 especially Drs. Mindy Goldsborough and James Kramer for supporting this work. This work was further supported 1004 by National Institutes of Health (NIH) Grants AI078859, AI074410, AI127351 -01, AI043894, and NS099029 1005 (to FK), R21DA050176 and R01NS099029 (to FK and LAL), and R33 CA206937 and R01AR068436 (to LAL).

The content is solely the responsibility of the authors and does not necessarily represent the official views of the 

Branscome et al. Fig. 7e e)