Differences in the immunophenotype of hematopoietic stem, progenitor cells of human immunodeficiency virus (HIV) exposed uninfected and umbilical cord blood

HEU Figure 1 shows the viSNE plots formed from immunophenotypic analysis of Lin-CD34+ haematopoietic stem and progenitor cells (HSPCs) from HIV exposed (HEU) and unexposed (HUU) samples. Umbilical cord blood was used and after removal of mononuclear cells, CD34+ magnetic isolation was performed. The CD34+ cells were stained with antibodies before and after 7 days of culture. This data was then analysed in an unsupervised way to determine differences. Expansion of cells occurred with cytokines only or with cytokines in addition to stemregenin-1 (SR1) is a small molecule expansion agent. On D0 almost no perceptible changes between the groups are observed. On D7, HUU samples have areas of bright CD90 expression while HEU samples have widespread dim expression of CD90. A flow cytometry self organising map (FlowSOM) was then run on the viSNE which shows the different metaclusters (MC) with MC 3 and 13 predominating in HUU samples and MC9 and 12 predominating in HEU samples. Each of these MC are given a name according to population nomenclature in the literature. Our immunophenotype panel was large and we thus added extended (Ext) in front of the names. These names are shown in HEU Table 1 where the percentages of the different populations are shown.In HEU Figure 2 box plots are portrayed which reveal the different MC's which are statistically different between HUU and HEU samples on D0, D7 with SR1 and D7 vehicle control (VC). After finding these differences, a CITRUS analysis was run to corroborate the differences in immunophenotype and this is shown in HEU Figure 3. HEU Figure 3 also shows a hierarchical plot and a volcano plot of the differences in gene expression on freshly isolated CD34+Lin- HSPCs between HUU and HEU samples. 271 differentially expressed genes were found although when we further analysed this, none of the genes played a significant role in any signaling pathway that could potentially affect cell function.<br>To further identify if any functional differences exist between HUU and HEU samples, colony forming unit analysis was performed. Freshly isolated CD34+Lin- cells were plated in a semi-solid medium called methocult for 14 days. After 14 days, the number and types of colonies produced were counted. The cells were also immunophenotyped. HEU Figure 4 shows that HUU and HEU samples produce the same number and type of colonies and on tSNE analysis of the immunophenotype, there are no differences in the MCs. Statistical analysis also revealed no differences in the MC's between the two groups.<br>In HEU Figure 5 we performed immunophenotypic analysis on the Lin-CD34- population. This population after culture has been shown to include both cells which have differentiated and become mature, and cells which may have downregulated their CD34 expression and are thus more immature. There is no nomenclature for this population and MC's were named as population (POP) 1-16. POP 5 is more prominent in HEU samples and statistical analysis showed many changes between HEU and HUU samples within the Lin-CD34- population. We are not sure whether these changes have clinical relevance.For the ARV component of the study, we isolated Lin-CD34+ haematopoietic stem and progenitor cells (HSPCs) from HIV unexposed uninfected neonates by magnetic isolation. We aimed to determine if antiretroviral drugs (ARVs) have any impact on HSPCs which could be a reason for any haematological abnormalities observed in HIV exposed but uninfected infants. The ARVs chosen were those which form part of the normal drugs pregnant mother sin South Africa take, namely Tenofovir disoproxil fumarate (TDF), lamivudine (3TC) and dolutegravir (DTG) in TLD fixed drug combination, and TDF, emtricitabine (FTC) and efavirenz (EFV) in TEE fixed drug combination. Based on what is known about cord blood to maternal serum ratios of the drugs mentioned, the maximum serum concentration was used as a starting point and concentrations above and below that used for the initial experiments. A lactate dehydrogenase (LDH) assay and a resazurin assay was used to determine cell cytotoxicity.<br>ARV Figure 1 shows the dose response curves after HSPCs were exposed to drugs for 24 hours and 7 days. In both instances the only drug found to have clear cytotoxicity which was dose dependent was EFV (decreasing cell viability and increasing LDH release with increasing concentrations). Dolutegravir shows increased cell number/viability in a 7-day culture at two concentrations which was reproducible in the 3 biological replicates.<br>The drugs were then combined in the FDC combinations mentioned above (TLD and TEE) and the same experiments performed. Here TEE was toxic to cells in a dose dependent manner and this was thought to likely be secondary to EFV. TLD showed an increase in cell number/viability again, and this was thought to be DTG, as with individual drugs experiments. These findings are shown in ARV Figure 2<br>A higher number of HSPCs were then cultured with drugs, to determine the impact of TEE and TLD on the number of cells obtained after 7 days of culture. Culture was in the presence of cytokines. Besides cell number, immunophenotype was also performed after 7 days to determine how drugs affect cell immunophenotype. ARV Figure 3 shows that HSPCs in the presence of TLD had higher cell counts on D7, followed by the vehicle control (VC) and then TEE. These differences were not statistically significant. a Uniform Manifold Approximation and Projection (UMAP) was run on the immunophenotype which showed multiple statistically significant differences between the drugs conditions and VC. These population metaclusters were gated and named according to known nomenclature. The names and frequencies are shown in ARV Table 1.A colony-forming unit experiment was also undertaken to determine if ARV presence has any impact on the formation of normal haematopoietic colonies. Drugs were cultured with freshly isolated CD34+Lin- HSPCs in methocult for 14 days. After 14 days the colonies were counted and immunophenotyping performed. ARV Figure 4 shows that in the presence of TLD, no erythroid colonies were found as compared to TEE and the VC. This was confirmed on the UMAP image as well as the back-gated Flow self-organising map (FlowSOM). The populations identified and the frequencies within each drug condition are shown in the table within Figure 4. ARV Figure 5 shows that the absence of erythroid colonies in TLD is due to a lack of a CD235a and CD71 co-expressing population. The statistically significant differences between TLD and VC are shown and the immunophenotypic differences between these significant populations is shown in heatmaps.<br><br>