Human microglia extensively reconstitute in humanized BLT mice with human interleukin-34 transgene and support HIV-1 brain infection

Humanized bone marrow-liver-thymic (hu-BLT) mice develop a functional immune system in periphery but have a limited reconstitution of human myeloid cells, especially microglia, in CNS. Further, whether bone marrow derived hematopoietic stem and progenitor cells (HSPCs) can enter the brain and differentiate into microglia in adults remains controversial. To close these gaps, in this study we unambiguously demonstrated that human microglia in CNS were extensively reconstituted in adult NOG mice with human interleukin-34 transgene (hIL34 Tg) from circulating CD34+ HSPCs but no in hu-BLT NOG mice, providing strong evidence that human CD34+ HSPCs can enter adult brain and differentiate into microglia in CNS in the presence of hIL34. Further, the human microglia in the CNS of hu-BLT-hIL34 NOG mice robustly supported HIV-1 infection reenforcing the notion that microglia are the most important target cells of HIV-1 in CNS and demonstrating its great potential as an in vivo model for studying HIV-1 pathogenesis and evaluating curative therapeutics in both periphery and CNS compartments.

Microglia，the resident macrophages in the central nervous system (CNS), are the key resident immune cells 49 to maintain neuronal homeostasis, defend against infections, and are associated with the pathogenesis of many 50 neurodegenerative diseases (1)(2)(3). The ontology of adult brain microglia has been debated for a long time. 51 The consensual view to date is that microglia in CNS is the seeding results of primitive hematomyeloid 52 precursor cells from yolk sac and aorta-gonad-mesonephros region in early embryo life and proliferation in 53 situ thereafter (4-8). However, multiple studies also showed that bone marrow derived cells can enter the 54 brain and differentiate into microglia in adults (9-11). 55 56 Humanized mice (hu-mice) with a human immune system have been extensively used in investigating the 57 ontology of immune cells, immunopathogenesis of human specific pathogens, and evaluating therapeutics as 58 preclinical small animal models(12-15). Hu-mice generated by engrafting human CD34+ hematopoietic stem 59 and progenitor cells (HSPCs) in neonatal life can reconstitute macrophages in the meninges and perivascular 60

Human brain tissues 114
To compare the human microglia from hu-BLT-hIL34 mice with humans, ethically sourced human autopsy 115 cerebral cortex tissues from a deidentified individual of HIV-1 negative with no registered medical 116 complications were obtained from the NIH Neuro BioBank (https://neurobiobank.nih.gov/ ). 117 118

HIV-1 infection and measurement of HIV-1 plasma viral loads 136
To investigate the infectivity of HIV-1 in parenchymal human microglial cells in the CNS, 4 mice (# 1703, 137 1705, 1708 and 1709) of the hu-BLT-hIL34 mice group and 5 mice (# 1720, 1723, 1724, 1726, and 1728) in the hu-BLT mice group were randomly selected (Table 1) and intraperitoneally inoculated with the 10 5 tissue 139 culture infectious dose 50 (TCID50) of macrophage-tropic HIV-1 ADA in 200μl volume (obtained through 140 the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH: HIV-1 ADA Virus from Dr. Howard 141 Gendelman). At 2-and 4-weeks post HIV-1 inoculation, HIV-1 plasma viral load (pVL) in copies/ml was 142 determined by real-time RT-PCR using our previously published protocol(24). Briefly, viral RNA was 143 extracted from the plasma using QIAamp ViralRNA minikit (Qiagen) as recommended by the manufacturer 144 and quantified using C1000 ThermalCycler and the CFX96 Real  (Fig. 1A). Whole brain was dissected out during 156 necropsy and sliced coronally into 5 parts at 4 mm interval using a young mouse brain slicer (Cat# 157 BSMYS001-1, Zivic Instruments, Pittsburgh, PA, USA). The brain tissues and other tissues including spinal 158 cord, spleen, lymph node, jejunum and ileum were collected and fixed in SafeFix TM II (Cat# 042600, Fisher 159 Scientific, USA) at room temperature for 6 hours and embedded in paraffin. 160 161

HIV-1 viral RNA detection using RNAscope in situ hybridization (ISH) 182
HIV-1 viral RNA (vRNA) in the brain tissues were detected using RNAscope ISH according to our previously 183 published protocol (26). Briefly, HIV-1 antisense probes of RNAscope® ISH probe-V-HIV1-clade B (Cat# was used as a negative control. All the reagents above were purchased from the Advanced Cell Diagnostics, 186 Inc. 187 188

Combined RNAscope ISH with IHCS 189
To determine the cell types of vRNA+ cells in the CNS, a combined RNAscope ISH and IHCS method was 190 used as reported (26). Briefly, after the completion of RNAscope ISH for HIV-1 vRNA and digitization of the 191 whole tissue section, the slides were soaked in xylene overnight to remove the coverslips and the tissue 192 sections were rehydrated and subjected to IHCS using the rabbit mAb to hIba-1 ((EPR6136-2 clone, Cat# 193 ab221933, 1:500; Abcam, USA)) as the primary antibody as described in the IHCS section above. Rabbit IgG 194 isotype control antibody was used as negative control. 195 196 Results 197 198 Human myeloid cell reconstitution in the CNS 199 1+ or hCD14+ positive cells using IHCS in the brain and spinal cord (Fig. 2 & 3). As shown in a representative 205 whole brain tissue section from the third coronary slice (Fig. 2C), hIba-1+ cells in the brain parenchyma were 206 extensively reconstituted across multiple regions of the brain. Fig. 2A and 2D respectively highlighted the 207 cerebral cortex and hippocampus boxed regions from the Fig. 2C; in turn, Fig. 2B and 2E respectively further 208 highlighted the boxed regions from the Fig. 2A and 2D at a higher magnification. The hIba-1+ cells are 209 numerous, morphologically ramified, distributed in brain parenchyma. Consistent with the extensive 210  , D). In turn, the blue and red boxed regions of the Fig A and B were further highlighted (B, E). The lower panel (F-H) shows a representative whole brain tissue section (G) of a hu-BLT-hIL34 mouse (#1708) that was stained immunohistochemically for hCD14+ cells (brown). The blue and red boxed regions was highlighted at a higher magnification (F, H). The hIba-1+ or hCD14+ cells are morphologically ramified and mainly distributed in brain parenchyma. reconstitution of parenchymal human myeloid cells in the CNS revealed by hIba-1+ cells, there were also 211 abundant hCD14+ myeloid cells as shown in a representative whole brain tissue section (2G) of a hu-BLT-212 hIL34 mouse (#1708). The blue and red boxed regions of cerebral cortex and hippocampus in the Fig. 2G  213 was respectively highlighted at a higher magnification in the Fig. 2F and 2H. The hCD14+ cells are again 214 morphologically ramified and mainly distributed in brain parenchyma. Consistent with the abundant 215 reconstitutions of human myeloid cells in the brain, we also observed the similar reconstitutions in the spinal 216 cord (Fig. 3). The upper panel of Fig. 3  hu-BLT-hIl34 mouse. We thus concluded that hu-BLT-hIL34 mice extensively reconstituted parenchymal 221 human myeloid cells in the CNS, including cerebral cortex, hippocampus, thalamus hypothalamus, striatum, 222 amygdala, spinal cord, and cerebellum (Supplemental Fig. 1). In contrast and as expected, there were 223 limited reconstituted human myeloid cells in the parenchyma of the CNS of the hu-BLT mice based on hIba-224 1 (Fig. 3G-K). As shown in a representative whole brain tissue section from the third coronal slice (Fig. 3I,  225 left upper), Iba-1+ cells were rare in all the parenchyma of cerebral brain (Fig 3I). Similarly, hCD14+ cells 226 were also rare in the parenchyma of hu-BLT mice (Fig. 3 E-F). We next compare the reconstitution of other 227 three types of nonparenchymal perivascular macrophages, meningeal macrophages, and choroid plexus 228 macrophages in the hu-BLT-hIL34 and hu-BLT mice. The subset of hIba-1+ or hCD14+ cells from hu-BLT-229 hIL34 mice also expressed hCD68 and hCD163. As indicated by hIba-1, hCD14, hCD163 expression and anatomic distribution, hu-BLT-hIL34 mice also have a better reconstitution of meningeal 231 macrophages (MM) and perivascular macrophages (PVM) (Supplemental Fig. 2). than hu-BLT mice (Fig 3  232

G-K). 233 234
The myeloid cells in brain parenchyma expressed microglia-specific marker TMEM119. 235  To distinguish parenchymal microglia from macrophages, we conducted IHCS using microglial-specific 236 marker hTMEM119 (31, 32). There were extensively reconstitutions of hTMEM119+ cells in the CNS 237 parenchyma of hu-BLT-hIL34 mice (Fig. 4). As shown in a representative whole brain tissue section from the 238 third coronal slice (Fig. 4C), hTMEM1191+ cells in the brain parenchyma were extensively reconstituted 239 across multiple regions of the brain. The Fig. 4A and 4D   In this study, we demonstrated that human microglia can be extensively reconstituted in CNS from circulating 264 human HSPCs in hu-BLT-hIL34 mice. We first used a battery of human myeloid cell markers, including hIBa-265 1, hCD14, hCD68 and hCD163, to evaluate human myeloid cell reconstitution in the CNS and found that 266 human myeloid cells were extensively reconstituted and primarily localized in the brain parenchyma (Fig 2  267 & 3). We then used a human microglial specific marker, hTMEM119, to validate these reconstituted human 268 myeloid cells in the brain parenchyma are mainly human microglia (Fig 4. A-F). Further, in comparison with 269 hTMEM119 + microglia in the cerebral cortex of a HIV-1 non-infected individual with no registered medical 270 complications (Fig 4. G-I), we found the frequency, distribution, and morphology of TMEM119+ human 271 microglia in hu-BLT-hIL34 mice are similar to this person. Our data thus support the notion that human 272 microglia at adults can be generated through human hematopoietic stem and progenitor cells (HSPC), which 273 is consistent with the previous reports that bone marrow derived cells can enter brain to different into 274 microglia at adults (9-11). In contrast, we did not observe human microglia reconstitution in the brain 275 parenchymal of hu-BLT mice. The hu-BLT-hIL34 and hu-BLT mice are genetically identical and also received 276 the same human donor transplant except the former had hIL34 knock-in, indicating that hIL34, a ligand of 277 the colony stimulating factor-1 receptor, play an important role in myeloid and microglial cells development 278 in CNS (19, 33). This study is unique in the following aspects. First human myeloid and microglia cells 279 reconstitution in the CNS is a clear-cut result in this chimeric mouse and human model. Second, the adult 280 mice engrafted with HSPCs extensively reconstituted human microglia in the CNS, to our knowledge this is 281 the first report in this regard. In addition to comparing the parenchymal human microglia between hu-BLT-282 hIL34 and hu-BLT mice, we also observed that hu-BLT-hIL34 mice had a much better reconstitution of 283 meningeal, perivascular, and choroid plexus macrophages than hu-BLT mice. We also would like to point that 284 both hu-BLT-hIL34 and hu-BLT mice received sublethal irradiation, whether this irradiation facilitated 285 HSPCs to gain an entry into the brain and whether without irradiation can also reconstitute the brain microglia 286 in adult hIl34-NOG mice remains to be investigated.

288
Despite the importance of microglial cells, as a resident macrophage, in host immune response to brain 289 infections and in the pathogenesis of neurodegenerative diseases, very limited small animal models are 290 available to recapitulate diseases pathogenesis associated with human microglia. To that end, we infected hu-291 BLT-hIL34 mice and found that reconstituted human microglia can support HIV-1 infection in the CNS.