Ready-to-use iPSC-derived microglia progenitors for the treatment of CNS disease in mouse models of neuropathic mucopolysaccharidoses – Nature Communications

Cell line generation

Human iPSC line generation and maintenance culture

The hiPSC lines 82 and 83 were derived by reprogramming peripheral blood mononuclear cells (PBMCs) to induced pluripotent stem cells. Peripheral blood was collected by the National Marrow Donor Program and donor informed consent was obtained (IRB protocol number Pro00025408). Erythroblasts were isolated and cultured from PBMCs according to Perriot et al.⁶⁹. Erythroblasts were electroporated with a synthetic self-replicating Venezuelan equine encephalitis (VEE) RNA replicon as described by Yoshioka and colleagues⁷⁰. The VEE replicon is a positive-sense, single-stranded RNA that does not utilize a DNA intermediate, so there is no potential for genomic integration and the absence of viral structural genes renders the RNA incapable of viral packaging and pathology. Human iPSC line 6 was obtained from Lonza⁷¹, donor informed consent for this line has been examined by NINDS and found to meet the highest standards.

hiPSCs were maintained using Essential 8 medium (Thermo Fisher Scientific, A26559-01) with recombinant laminin (Biolamina, LN521-05) or Vitronectin (Thermo Fisher Scientific, A14700) as the hiPSC attachment substrate. Cell passaging was performed using EDTA as a non-enzymatic method (Thermo Fisher Scientific, AM9260G) or Accutase (Thermo Fisher Scientific, 00-4555-56). Dissociated cells were plated onto Laminin 521 coated tissue culture vessels in Essential 8 medium supplemented with 10 µM of Y-27632 (Bio-Techne, TB1254-GMP) for the first 24 hr. Parental and clonally-derived gene-edited hiPSC lines were screened for normal karyotypes by Cell Line Genetics (Madison, WI). All hiPSC lines were routinely analyzed for appropriate expression of the pluripotency markers NANOG, SOX2, OCT4, SSEA-3, SSEA-4, and TRA-1-60 (BD, 560873, 561506, 560329, 560236, 563119, and 561153 respectively). Pluripotency and karyotype information is found in Supplementary Fig. 13.

hiPSC knockout cell line generation

hiPSC lines lacking IDUA, GUSB, NAGLU, and SGSH were generated to test the ability of MG01 to in vitro cross-correct this enzyme deficiency. CRISPR/Cas9 single guide RNAs (sgRNAs) targeting a coding exon of each of these four genes were designed and tested. The sequences of these sgRNAs are listed in Table 1. A vial of the parental hiPSC line 82 was thawed and transfected with individual Cas9:sgRNA ribonucleoprotein (RNP) complex targeting each of the four genes to generate functional knockouts. The transfected cells were plated onto Vitronectin-coated tissue culture vessels in Essential 8 medium containing Y-27632 (10 µM). Clones were manually isolated by picking and then screened for bi-allelic gene KO using PCR primers flanking the sgRNA target site followed by Sanger sequencing. PCR primers used for genotyping are listed in Table 2. Multiple clones of mono- and bi-allelic KO genotypes were identified for each target gene, but only bi-allelic KO clones were selected for MG01 differentiation. Final genotypes for each KO line are shown under the “Genotype” column in Table 1.

MG01 differentiation

hiPSCs were plated onto Vitronectin (Thermo Scientific, A14700) at 1.0 × 10⁴ cells/cm² in Essential 8 medium (Thermo Scientific, A1517001) containing 10 µM of Y-27632 (Tocris, 1254) for 24 hours. hiPSCs were cultured for an additional 2 days in Essential 8 medium with daily medium changes before being induced by Essential 6 (Thermo Scientific, A1516401) medium supplemented with 80 ng/mL of BMP-4 (R&D Systems, 314E-GMP-050). BMP-4 induction continued for 4 days with daily medium changes before the cultures were changed to StemPro-34 SFM medium (Thermo Scientific, 10639011) (containing 1X GlutaMAX Thermo Scientific, 35050061) supplemented with 100 ng/mL of SCF (R&D Systems, 255B-GMP-050), 80 ng/mL of VEGF (R&D Systems, 293-GMP-050) and 25 ng/mL of bFGF (R&D Systems, 233-GMP-025) for 2 days with daily medium changes. On day 6 and day 8, the cells were cultured with StemPro-34 SFM medium containing 50 ng/mL of SCF, 50 ng/mL of IL-3 (R&D Systems, 203-GMP-050), 50 ng/mL of M-CSF (R&D Systems, 216-GMP-500), 50 ng/mL of FLT3L (R&D Systems, 308E-GMP-050) and 5 ng/mL of TPO (R&D Systems, 288-TPE-050). Starting on day 10, cells from the supernatant fraction were pelleted, resuspended in the same fresh medium as days 6 and 8, and placed back into their respective vessel. Beginning on day 14, cells in the supernatant fraction were pelleted, resuspended in StemPro-34 SFM medium containing 50 ng/mL of FLT3L, 50 ng/mL of M-CSF and 25 ng/mL of GM-CSF (R&D Systems, 215-GMP-050) and placed back into their respective vessel. Day 14 medium change was repeated every other day until the cells in the supernatant reached a concentration of >1.0 × 10⁶ live cells per mL. Beginning on day 18 and every other day after, cell counts were performed prior to pelleting the cells in the supernatant for medium exchange to determine the harvest date.

Cryopreservation and thawing of MG01

MG01 were isolated and counted on the NucleoCounter NC-200 (Chemometec) prior to freezing. Cells were resuspended in either BamBanker (Wako Chemicals, 30214681)) or STEM-CELLBANKER GMP grade (amsbio, 11924) freezing medium and transferred to cryogenic vials (Thermo Scientific). Cryogenic vials with cells were frozen with a controlled rate freezer (CBS CRF2101). Cells were transferred to liquid nitrogen (vapor phase) for long-term storage. To thaw MG01, cryogenic vials were transferred to a 37 °C water bath for ~2 minutes until a small ice crystal remained, transferred to a centrifuge tube, and quenched with 1 mL of either RPMI-1640 (Thermo Scientific, 11-875-101) or StemPro-34 SFM medium in a drop-wise fashion. The cells were counted on the NC-200 and quenched with an additional 3 mL of its respective medium. Cells were centrifuged at 250 g for 5 minutes and resuspended in the appropriate assay or culture medium.

Replating and polarization of MG01

MG01 were plated at 1.0 × 10⁵ cells per well of a 96-well tissue culture treated plate in 50 µL of RPMI-1640 (Thermo Scientific, 11-875-101) containing 1X GlutaMAX supplemented with 100 ng/mL of IL-34 (R&D Systems, 5265-IL-010/CF) and 10 ng/mL of GM-CSF (maturation medium). For M0-like microglia, an additional 50 µL of maturation medium was added directly on top of the cells. For M1-like microglia, an additional 50 µL of maturation medium supplemented with 200 ng/mL of LPS (Sigma-Aldrich, L4391) and 200 ng/mL of IFNγ (R&D Systems, 285-GMP) was added directly on top of the cells, unless otherwise stated in the figure legend. For extended microglia culture, MG01 was replated as previously described for 6 days or 11 days with media exchanges at 48 hour intervals. Cultures were polarized as previously described 48 hours prior to E.coli phagocytosis challenge. After addition of pHrodo Red E.coli bioparticles (Thermo Fisher Scientific, P35361), wells were imaged every hour for 24 hours. All cell lines tested behaved similarly and the data depicted is representative of all hiPSC-derived MG01.

Cytokine release assay (FRET)

Spent medium collected from 48 hour polarized microglia were centrifuged at 250 g for 5 minutes and the supernatants were frozen at −80 °C. Supernatants were thawed at room temperature and 16 µL were transferred to small-volume 96 well plates for the addition of donor and acceptor antibodies from either the IL-6, TNFα, or CXCL10 kit (Cisbio, 62HIL06PEG, 62HTNFAPEG, 62HCX10PEG). Following the manufacturer’s instructions, samples were incubated for 2–24 hours at room temperature and processed along with standards, and analyzed on a Clariostar microplate reader (BMG Labtech) using the recommended parameters from Cisbio. Data was interpreted by calculating the ratio of the acceptor and donor emission signals and determining the delta ratio (ratio standard or sample–ratio standard 0). Supplementary Fig. 1C does not include CXCL10 data as it could not be collected due to technical failure.

Flow cytometry analysis

MG01 cells were thawed in 2 mL maturation medium as previously described and a cell count was performed on the NucleoCounter NC-200 (Chemometec). Based on the live cell count per mL, 5.0 × 10⁵ cells were transferred to a 5 mL FACS tube (Falcon). Aliquoted samples of cells were prepared and placed into separate tubes for unstained and viability dye controls. Additional maturation medium was added to the cells at a 1:5 dilution to further dilute the cryopreservation medium. The cells were pelleted by centrifugation at 250 g for 5 minutes, the supernatant was carefully removed, and the cells were resuspended in 100 µL of surface marker staining cocktail shown in Table 3. The unstained sample was resuspended in 100 µL of 1X DPBS (Thermo Scientific, 14190250) and viability dye (Thermo Scientific, NC0476349) control was resuspended in 100 µL of 1X DPBS containing 0.33 µL of the viability dye. The samples were incubated at 4 °C for 30 minutes, washed with 2 mL of 1X DPBS, and pelleted by centrifugation at 200 g for 5 minutes. Samples were resuspended in 200 µL of FACS buffer (Table 3) and then analyzed on the CytoFLEX LX (Beckman Coulter). FCS files were exported and analyzed using FlowJo v10.10 software. The representative gating strategy is included in Supplementary Fig. 14.

Phagocytosis of E. coli pHrodo bioparticles

MG01 were plated at 1.0 × 10⁵ cells per well of a 96-well tissue culture plate and then polarized to either M0- or M1-like microglia. After 2 days, medium change was performed by including pHrodo Red E. coli Bioparticles (Thermo Scientific, P35361, prepared according to manufacturer’s instructions) in both M0- and M1-like microglia maturation medium (see polarization section) at a final concentration of 45.45 µg/mL per well. Cells on the plate were placed into the IncuCyte S3 (Sartorius) within a 37 °C incubator with 5% CO₂ and imaged every 2 hours for phagocytic activity. The orange signal is detected from the background, partitioned according to edge sensitivity parameters, filtered out if <100 μm² in the area, and then the average OCU of the object will be multiplied by the area of the orange object. Total integrated intensity is the total sum of the objects’ fluorescent intensity after background subtraction per image and it is calculated at each timepoint on each well then averaged within the condition. One well with M0 medium including pHrodo Red E. coli Bioparticles only (no cells), and a single well with 10 × 10³ MG01 with M0 medium alone (no pHrodo) were also imaged as controls. All cell lines tested behaved similarly and the data depicted is representative of all hiPSC-derived MG01.

Immunofluorescent staining for imaging

Cells were fixed with cold 4% paraformaldehyde (Thermo Scientific, AAJ61899AK) for 10 minutes at room temperature (RT). After fixation, cells were washed with 3 rounds of 1X DPBS (Thermo Scientific, 14190250) for 5 minutes at RT before being permeabilized with three washes of 1X DPBS containing 0.2% Triton X-100 (Sigma-Aldrich, T8787) for 10 minutes each at RT. Next, 10% normal donkey serum (Jackson ImmunoResearch, 017-000-121) in 1X DPBS was added to the cells for 1 hour at RT, and then cells were stained with primary antibody cocktail (Table 4) in a blocking solution at 4 °C overnight. Cells underwent 3 rounds of washes with 1X DPBS containing 0.2% Triton X-100 (DPBS-T) for 10 minutes each at RT the next morning prior to adding the secondary antibody cocktail in 1X DPBS at RT for 1 hour. An additional round of three 10-minute DPBS-T washes preceded a 20-minute incubation with DAPI (Invitrogen, D21490) in 1X DPBS at RT. Lastly, cells were washed with 1X DPBS and stored at 4 °C.

Quantification of LAMP1 staining was performed utilizing QuPath 0.4.3⁷² utilizing the Cell Detection workflow using DAPI for nuclei detection, with a cell expansion radius of 10um, and analyzed with Kruskal-Wallis test followed by Dunn’s multiple comparisons test.

In-vitro metabolic cross-correction

Wild-type and enzyme-deficient (MG-KO) microglia were thawed separately in the maturation medium. Wild-type microglia were plated in 6 well tissue culture-treated plates at 1.0 × 10⁶ in 2 mL of maturation medium. In a separate 6-well tissue culture treated plate holding 0.4 µM porous transparent polyester membrane inserts (Falcon, 353090), MG-KO were plated at 5.0 × 10⁵ in 1 mL of maturation medium (2 mL of maturation medium was added to the well below to prevent the membrane from drying out). After 48 hours, MG-KO on inserts were transferred and placed on top of wild-type wells to begin co-culture. Medium change was performed at the start of co-culture and every other day for 10-14 days. Standalone MG-KOs on inserts were cultured alone in parallel, as a diseased control. To measure metabolic cross-correction wild-type and enzyme-deficient microglia were harvested separately from the insert and wells, lysed in 1% Triton X-100, and evaluated for intracellular enzymatic activity and total GAG content. For the conditioned media experiments, thawed MG01 and MG-KO (IDUA) were plated into separate wells of an 8-well chamber slide at 3.0 × 105 in 400 uL of maturation medium. Beginning on day 0, and every other day onwards, till day 10, the spent medium was collected from MG01 samples and replaced with 400uL of fresh maturation medium. MG01 conditioned medium was prepared by centrifuging the MG01 spent medium at 250 x g to remove any cellular debris. To condition MG-KO (IDUA^−/−), MG-KO spent medium was replaced with 400 uL of the MG01 conditioned medium (supernatant) containing 100 ng/mL of IL-34 and 10 ng/mL of GM-CSF during each medium exchange, for a duration of 10 days.

Neuronal differentiation and co-culture

Human iPSCs were differentiated into cortical neuronal progenitors as reported by Ciceri et al.⁷³ On day 20 neural progenitor cells (NPCs) were dissociated to single cells using Accutase and were resuspended in STEM-CELLBANKER (Amsbio, 11924) freezing medium and transferred to cryogenic vials. Cryogenic vials with cells were frozen with a controlled rate freezer (CBS CRF2101). Cells were transferred to liquid nitrogen (vapor phase) for long-term storage. 5.0 × 10⁵ NPCs were plated into separate 6-well tissue culture-treated wells holding 0.4 µM porous transparent polyester membrane inserts (Falcon, 353090). Plated NPCs were cultured in Neurobasal complete media supplemented 20 ng/mL of BDNF (R&D Systems,11166-BD-050), 10 ng/mL of GDNF (R&D Systems, 212-GD-050/CF), 0.5 mM of cAMP (Sigma, D0627), 0.2 mM ascorbic acid (Sigma, A4403) and 10uM DAPT (2634, Tocris) for 10 days with media exchanges every 3 days. Wild-type microglia were plated in 6 well tissue culture-treated plates at 1.0 ×10⁶ in 2 mL of maturation medium. Day 30 NPCs in transwells were transitioned to BrainPhys media (Stem Cell Technologies, 05790) containing N2, BDNF, GDNF, asorbic acid, IL-34, and GM-CSF and placed on top of wild-type microglia wells. Medium change was performed at the start of co-culture and every other day for 10 days. Standalone NPC-KOs on inserts were cultured alone in parallel, as controls.

Dose formulation for animal transplantation

MG01 were thawed and transferred to a centrifuge tube. Before performing a cell count on the NucleoCounter (Chemometec), 1 mL of StemPro-34 SFM medium supplemented with 50 ng/mL of FLT3L, 50 ng/mL of M-CSF, and 25 ng/mL of GM-CSF (SP34-d14) was added directly on top in a drop-wise fashion. An additional 3 mL of SP34-d14 medium was added on top to further dilute the freezing medium. Cells were centrifuged at 250 g for 5 minutes at room temperature and the supernatant was aspirated. The cell pellet volume was determined using a 20–200 µL single channel pipette and transferred to a 1.5 mL microcentrifuge tube. Based on the live cell count per mL, the cell pellet was diluted with additional StemPro-34 SFM-d14 medium for a final formulation of 7 × 10⁴ (high-dose) or 2 × 10⁴ (low-dose) live cells per µL. The prepared cells were placed on ice until ready for transplantation.

Enzymatic activity assay

To determine enzyme activity, cells from the membrane insert were lysed in 1% Triton-X 100 containing protease inhibitors followed by assessment of enzyme activity and GAG levels. The method was adapted from Ou et al., 2014⁷⁴. Enzyme catalytic activity for each lysosomal enzyme was determined by quantifying 4-Methylumbelliferone (4-MU), the fluorescent moiety produced after cleavage of an artificial fluorogenic substrate that is unique to each enzyme (Table 5). The resulting fluorescence of cleaved substrate was read on the CLARIOstar (MARS software v3.31) with an excitation setting at 355 nm and emission at 460 nm. Enzyme levels and activity were interpolated using a 4-MU standard curve. For animal experiments, frozen tissue was homogenized using 1% Triton X-100 in HBSS (Cytiva, SH30588.01) with Halt protease inhibitor (Thermo Scientific, 87786). A BCA protein assay was performed on the resulting lysates to estimate the amount of protein isolated. Up to 50 µg of total protein was used to measure enzyme activity and GAG levels from brain lysates, spinal cord lysates, and serum. For CSF, 1 µl of the sample was used for the assay and results were post-normalized to total protein concentration. Enzyme activity in urine was measured using 5 μL upon thaw. In parallel, creatinine levels were quantified to normalize urine enzymatic activity. Values before normalization to vehicle or wildtype controls are included in Supplementary Table 1.

Enzymatic activity for Line 6 after engraftment studies was assayed with a first generation of the IDUA enzyme activity assay, which needed to be further optimized. Prior to optimization, the strong signal from the fluorogenic compound used to measure IDUA activity showed some background, especially at higher enzyme concentrations. We thus re-optimized the assay to measure the lower enzyme levels expected in our grafted animals. As such, a direct comparison between Line 6 and Lines 82-83 is not possible.

Glycosaminoglycan (GAG) comprised of; dermatan sulfate (DS), heparan sulfate (HS), keratan sulfate (KS), chondroitin sulfate (CS), hyaluronic acid (HA).

GAG quantification

Frozen tissue was homogenized as described above (refer to Enzymatic Activity Assay section). Quantification of GAGs from the various cell lysates was performed using a cationic dye 1,9 dimethylmethelene blue (DMB) (Sigma Aldrich, 341088-1G) which binds to highly charged sulfated GAGs. The resulting absorbance at 525 nm is proportional to the concentration of GAGs in the sample. Values were interpolated using a GAG standard curve. GAG levels were measured using 5 μL of urine upon thaw. In parallel, creatinine levels were quantified to normalize urine GAGs. Values before normalization to vehicle or wildtype controls are included in Supplementary Table 1.

SGSH automated western

Samples were homogenized as described above (refer to Enzymatic Activity Assay section), and a BCA protein assay was performed to estimate the total protein concentration. 5ug of total protein (or a maximum of 4uL) was loaded into each Jess Automated Western reaction. A standard curve was also generated using recombinant human SGSH (Bio-Techne, 8380-SU). A 12-230 kDa separation cartridge was loaded with samples, standards, anti-SGSH antibody (20ug/mL, Bio-Techne MAB83801), and anti-mouse HRP secondary (Bio-Techne 042-205). The instrument was run according to manufacturer instructions, and the resulting peak heights that corresponded to the molecular weight of SGSH (in the electropherogram) were used to generate a standard curve and interpolate results.

LAMP1 and LAMP2 quantification

To study modulation of LAMP1 and LAMP2 levels, MG-KO, MG01, and cortical neurons were homogenized as described above (refer to Enzymatic Activity Assay section), and a BCA protein assay was performed to estimate the total protein concentration. 0.5 µg of total protein was diluted in the sample diluents provided by each kit, and LAMP1 (Abcam, ab302755) and LAMP2 (Abcam, ab302757) ELISAs were performed according to manufacturer protocols.

Animals and husbandry

Experiments were performed with male and female mice (P21 and P60-90 old; n = 6–8/age group/genotype) with comparative groups (NOD.Cg-Prkdc^scid Idua^tm1Clk/J [Idua ^−/−] versus their wild-type [Idua^+/+] littermates) (Strain #:004083, Jackson Laboratory) and NOD.Cg-Prkdc^scid Gusb^mps/SndsJ [Gusb^−/−] versus their wild-type [Gusb^+/+] littermates (Jackson Laboratory). For engraftment studies, male and female mice NSG-Q (NOD.Cg-Prkdc^scidIl2rg^tm1WjlTg(CMV-IL3,CSF2,KITLG)1Eav/MloySzj; Strain #013062, Jackson Laboratory, crossed with NOD.Cg-Prkdc^scidIl2rg^tm1WjlTg(CSF1)3Sz/SzJ; Strain #028654, Jackson Laboratory; P7, P14 and P21; n = 1-2/age group/sex) were used. Animals were maintained on a 12 h light-dark cycle (lights on at 6:00 am) at 22–25 °C with ad libitum access to food and water. Mice were housed 2–5 per cage. All experiments were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee under protocol number 2021-BRT-8.

Stereotaxic surgery

Briefly, 30 minutes prior to surgery, adult animals (P21 and older) were injected subcutaneously with 4 mg/kg Meloxicam SR (Putney, RXMELOXICAM-INJ), and all animals were anaesthetized by inhalation of ~2% isoflurane (Attane, RXISO-250) in oxygen throughout the entire surgical procedure. Anaesthetized animals were positioned in a stereotaxic frame (Kopf instruments) and 10 µL of either MG01 low dose (0.4×10⁶ cells/brain), MG01 high dose (1.4 × 10⁶ cells/brain), or vehicle (StemPro-34 SFM medium containing 50 ng/mL of FLT3L, 50 ng/mL of M-CSF, and 25 ng/mL of GM-CSF; 0 cells/brain) were injected bilaterally with a blunt, 27-gauge needle attached to a 10 µL Model 701 RN Hamilton syringe into the lateral ventricles using the following coordinates from bregma:

1. i.

   P60 & P90 mice: AP −0.3, ML+/− 1.4, DV 2.34 at a 10° angle

2. ii.

   P14 & P21 mice: AP + 0.1, ML+/− 1.4, DV −2.34 at a 10° angle

3. iii.

   P7 mice: AP + 0.1, ML+/− 1.4, DV −2 at a 10° angle

Hindlimb clasping Test

Behavioral testing was performed at 1, 5, and/or 8 months post-surgery, depending on the cohort. Briefly, animals were removed from cage and suspended by their tail for ~10-15 seconds and their abilities of hindlimb clasping were monitored via video recording. Hindlimb clasping was scored on a scale from 0-3 based on the following:

1. i.

   Score 0: Both hindlimbs were splayed outward and away from the abdomen

2. ii.

   Score 1: One hindlimb is retracted toward the abdomen for more than 50% of the time

3. iii.

   Score 2: Both hindlimbs are partially retracted toward the abdomen for more than 50% of the time

4. iv.

   Score 3: Both hindlimbs are entirely retracted and touching the abdomen for more than 50% of the time

Tissue collection and analysis

For in life analysis, blood was collected via submandibular bleed (50 ul) one day prior to surgery. Post surgery, blood (collected via submandibular bleed; 50 ul) and urine (collected via rubbing abdomen above bladder; 20 ul) samples were collected 2 weeks and then every month before termination. Terminal tissue collection was performed 1-, 5-, or 8-months post MG01 transplantation, depending on the cohort. Briefly, animals were anaesthetized by inhalation of ~2% isoflurane in oxygen and a small incision was made along the back of the neck to expose the muscle tissue. Next, muscle tissue was carefully removed to expose the cisterna magna and a 20 µm glass pulled pipette was inserted into the cisterna magna to collect ~3–8 µl of cerebrospinal fluid (CSF). Following CSF collection, blood was collected via intracardial puncture, urine was extracted from the bladder and then animals were perfused transcardially with cold phosphate buffered saline (PBS). Brains were dissected into two hemispheres and the left hemisphere was placed in cold 4% paraformaldehyde (PFA) for 48 h for histological analysis and the right hemisphere was flash frozen for biochemical analysis. One spinal column was collected from each group and placed in 4% PFA for 48 h for histological analysis and the remaining spinal cords were extruded via PBS flush and flash frozen for biochemical analysis. For serum extraction, blood was allowed to coagulate at RT for 1 hr and then centrifuged at 1300 g for 15 min at 4 °C. Clear supernatant, containing serum, was carefully removed to a new tube and flash frozen for biochemical analysis. For brain region specificity of enzymatic activity and substrate accumulation, the right hemisphere was placed into a brain block and cut at 1 mm thick sections coronally. Each region was removed with a 2 mm tissue puncher, and 2 punches per region were pooled for biochemical analysis.

Histology and immunofluorescence

Harvested tissues for histological evaluation were fixed in 4% PFA for 48 h. Left brain hemispheres were transferred into 20% sucrose (MP Biomedicals, 0219474705) in PBS for 24 h and then into 30% sucrose in PBS for 24 h. Brains were then embedded in OCT (Sakura, 4583)/30% sucrose solution (1:1) and snap-frozen in an ethanol bath, over dry ice. 20 µm thick sections were then obtained in a sagittal orientation using a Cryostat (LEICA; CM3050 S) and mounted onto positively charged microscope slides (Electron Microscopy Sciences, 71873-02). The slides were allowed to air dry for ~1 h and then stored at −80 °C. Spinal columns were washed in PBS and then the spinal cord was dissected from the vertebral column. Extracted spinal cords followed the same sucrose gradient submersion as described above for brain hemispheres. Following the last sucrose solution, spinal cords were dissected into cervical, thoracic, and lumbar regions, embedded in OCT solution, and stored at −80 °C. Spinal cords were sectioned into serial 20 µm slices on the Cryostat, mounted onto positively charged slides, and stored at −80 °C.

For immunofluorescence staining, slides were incubated at RT for 30 m, permeabilized with 0.2% Triton-X in PBS (PBS-T), blocked for 1 h with 10% donkey serum in PBS-T and incubated with primary antibodies (Table 4) in blocking solution overnight at 4 ^oC. The next day, slides were washed with PBS-T 3 times and incubated with secondary antibodies (Table 4) for 1 h at RT. Slides were then washed 3 times with PBS-T, incubated with 300 nM DAPI in PBS for 30 m, and mounted with Fluorsave™ Reagent (MilliporeSigma, 345789) for imaging. Slides were imaged in an Axioscan 7 microscope (Zeiss).

For DAB, sections were processed the same way as above with the addition of heat-induced antigen (epitope) retrieval (HIER) step for 30 minutes at 95 °C in 1x Citrate Buffer (Abcam, ab64214), permeabilization using 0.8% PBS-T, and utilizing a biotinylated secondary followed by PBS-T washing steps and streptavidin-DAB incubation for 30 m. After, DAB was developed utilizing a DAB peroxidase substrate kit (Fisher Scientific, SK-4100). Slides were imaged in an Axioscan 7 microscope (Zeiss).

hNA+ cell count statistical analysis

Quantification of MG01 was performed by cryosectioning of formalin fixed brain hemispheres followed by immunohistochemistry for the human marker hNA or ku80. Sagittal cryosections 200-400um apart spanning the whole brain of every animal were selected to check for hNA immunoreactivity after manual assessment of image and section quality. Sections were imaged using a Zeiss Axioscan 7 Slide Scanner and hNA counts were performed using an automated workflow based on pixel classification using the Intellesis platform in Zen Blue 3.3 (Zeiss).

Ki67+ colocalization analysis

Quantification of ki67+/ku80+ cells was done manually using the Cell Counter plugin in ImageJ or Fiji⁷⁵. Three comparable sections per animal were selected to check for ki67/ku80 immunoreactivity after manual assessment of image and section quality. All ku80 cells detected were counted and the proportion of ki67-positive cells in that population was calculated. All data points are the average of the percentage per section per animal.

Statistical analysis

All statistical analyses were performed using Prism 9.4.1 (GraphPad Software LLC). In all graphs: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. For all IDUA levels, GAG quantification, and behavior, data points represent biological replicates.

For enzymatic activity and GAG readouts, data was tested for normality using the Shapiro-Wilk normality test. If normally distributed, it was analyzed using Prism with a one-way ANOVA using Dunnet’s multiple comparison test, comparing all groups to the vehicle group. If not normally distributed, it was analyzed with Kruskal-Wallis (KW) test followed by Dunn’s multiple comparisons test. Data is represented as mean and SEM.

For Fig. 5, there was a significant effect of treatment on GAG levels in the 4-month MPS I adult cohort treated with Line 6 (F_2,8 = 96.76, p < 0.0001). Similarly, a treatment effect on GAG levels for the juvenile MPS I cohort brain lysates at 1 M (F _{(3, 21)} = 1107, P < 0.0001), 5 M (F _{(3, 16)} = 270.3, p < 0.0001) and 8 M (KW: H = 16.02, p < 0.001); in spinal cords at 1 M (F _{(3, 13)} = 306.5, P < 0.0001), 5 M (F _{(3, 12)} = 151.8, p < 0.0001), 8 M (F_{(3, 11)} = 26.63, p < 0.0001); and CSF at 1 M (F _{(3, 21)} = 179.5, P < 0.0001) and 5 M (F _{(3, 16)} = 304.7, p < 0.0001). There was a treatment effect on GAG levels for the Adult MPS I cohort brain lysates at 1 M (F _{(3, 20)} = 182.6, p < 0.0001) and 5 M (F _{(3, 18)} = 140.3, p < 0.0001); in spinal cord at 1 M (F _{(3, 20)} = 182.6, p < 0.0001) and 5 M (F _{(3, 14)} = 72.50, p < 0.0001); and CSF at 1 M (F _{(3, 20)} = 41.89, p < 0.0001). In the MPS VII model, there was a significant effect on adult cohort brain lysates at 1 M (KW: H = 13.66, p < 0.0001) and 5 M (F _{(2, 23)} = 252.4, p < 0.0001), spinal cord at 1 M (F _{(2, 15)} = 204.2, p < 0.0001) and 5 M (F _{(3, 20)} = 76.32, p < 0.0001). In the juvenile cohort, there was a significant effect at 1 M on brain lysates (F _{(3, 21)} = 76.21, p < 0.0001) and spinal cords (F _{(3, 17)} = 31.38, p < 0.0001). p values from post-hoc tests when compared to the vehicle group are displayed on the corresponding figure.

For Supplementary Fig. 9, there was a significant effect of treatment on IDUA levels in the juvenile MPS I cohort brain lysates at 1 M (F _{(3, 21)} = 559.7, p < 0.0001), 5 M (KW: H = 17.86, p = 0.0005)) and 8 M(F _{(3, 15)} = 764.9, p < 0.0001); spinal cords at 1 M (F _{(3, 21)} = 559.7, p < 0.0001), 5 M (F _{(3, 12)} = 3898, p < 0.0001) and 8 M (F _{(3, 11)} = 1569, p < 0.0001); and CSF at 1 M (F _{(3, 21)} = 179.5, p < 0.0001), 5 M (F _{(3, 16)} = 575.9, p < 0.0001) and 8 M (F _{(3, 14)} = 116.4, p < 0.0001). In the adult MPS I cohort, there was a significant effect in brain lysates at 1 M (F _{(3, 20)} = 7921, p < 0.0001) and 5 M (F _{(3, 18)} = 2256, p < 0.0001), and in CSF at 1 M (KW: H = 20.42, p = 0.0001) and 5 M (F _{(3, 16)} = 304.7, p < 0.0001). p values from post-hoc tests when compared to the vehicle group are displayed on the corresponding figure.

For hindlimb clasping, a minimum of 5 animals per group were used across all studies. As the scoring from hindlimb clasping is not normally distributed, data were analyzed using R and the nonparametric Kruskal-Wallis test and Conover-Iman post hoc tests were used if significant main effects were observed via Kruskal-Wallis. All data are represented as median.

For Fig. 7, there was a significant effect of treatment in hindlimb clasping score in MPS I animals at 5 M (chi-squared = 9.7119, df = 3, p = 0.02) and 8 M (chi-squared = 8.8107, df = 3, p = 0.03) post-treatment. In MPS VII animals, there was a significant effect on treatment in hindlimb clasping score in the adult cohort at 1 M (chi-squared = 13.6137, df = 2, p = 0), 2 M (chi-squared = 22.6845, df=2, p = 0), and 5 M (chi-squared = 16.733, df = 2, p = 0) post-treatment; and in the juvenile cohort at 1 M (chi-squared=8.0653, df = 2, p = 0.02) post-treatment. p values from post-hoc tests are displayed on the corresponding figure.

For urine and serum samples, data was analyzed using Prism using a 2-way ANOVA in a mixed-effects analysis followed by Dunnet’s multiple comparison test, comparing all groups to the vehicle group. Data is represented as mean and SEM.

For Fig. 6, a statistically significant interaction between the effects on time and treatment on GAG levels in urine of the adult MPS I 1 M (F _{(3, 20)} = 396.5, p < 0.0001) and 5 M (F _{(12, 75)} = 274.4, p < 0.0001) cohorts; in the MPS I juvenile 1 M (F _{(3, 20)} = 85.57, p < 0.0001), 5 M (F _{(12, 47)} = 29.64, p < 0.0001) and 8 M (F _{(18, 71)} = 15.89, p < 0.0001) cohort; in the MPS VII adult 1 M (F _{(4, 36)} = 21.98, p < 0.0001) and 5 M (F _{(12, 107)} = 6.824, p < 0.0001) cohort. In the MPS VII juvenile 1 M cohort there was a significant effect of treatment (F _{(3, 32)} = 176.1, p < 0.0001) on GAG levels in urine. p values from post-hoc tests are displayed on the corresponding figure.

For Supplementary Fig. 11, there was a significant effect of treatment on IDUA levels of the MPS I adult cohort at 1 M (F _{(3, 20)} = 932.3, p < 0.0001). There a statistically significant interaction between the effects on time and treatment on IDUA levels in serum of the adult MPS I 5 M (F _{(12, 86)} = 10.84, p < 0.0001) cohort; and in the juvenile 1 M (F _{(6, 42)} = 16.09, p < 0.0001), 5 M (F _{(18, 88)} = 49.82, p < 0.0001) and 8 M (F _{(21, 105)} = 38.64, p < 0.0001) cohorts. There was a significant effect of treatment on GAG levels in the heart (F _{(3, 16)} = 964.9, p < 0.0001) and liver (F _{(3, 16)} = 1104, p < 0.0001) of the 5 M MPS I cohort and in the liver of the 5 M MPS VII cohort liver (F _{(2, 16)} = 24.18, p < 0.0001). p values from post-hoc tests are displayed on the corresponding figure.

Digital droplet polymerase chain reaction (ddPCR)

ddPCR assay and analysis was conducted at a contract research organization (Charles River, SF). Briefly, the ddPCR assay was a singleplex assay designed to detect hTERT DNA, consisting of oligonucleotide primers and probe mix containing a TaqMan™ hTERT Control Assay (Thermo Fisher) designed to amplify the hTERT sequence in human gDNA. 18 mice (9 males and 9 females) were dosed via bilateral intracerebroventricular (ICV) injection with MG01 or formulation vehicle. The samples were collected 1 month after dosing from mouse whole blood, plasma, brain, spinal cord, liver, spleen, lungs, and kidney fluids and tissues and DNA was isolated. Prepared dosed samples from shedding matrices and biodistribution tissues were analyzed via ddPCR in duplicate. The fluid and tissue samples were analyzed in a BioRad QX ONE ddPCR at a final concentration of 50 ng/rxn of gDNA. All ddPCR analysis contained 2 sets of controls, a no template control, and a negative control.

Bulk RNA sequencing

RNA was isolated from micro-dissected fixed histology sections using the RNeasy FFPE kit (Qiagen) and analyzed on an Agilent 2100 Bioanalyzer using Agilent 6000 RNA Nano Kit (Agilent Technologies). Samples included in analysis are post-thaw MG01 (Line 6) and isolated tissue from an NSG-Q mouse 5.5 months after intracranial transplantation with Line 6-derived MG01 (Coordinates: 7° angle- ML + 2, AP −2.2, DV −2 > −1.8, −0.95). After RNA extraction, all samples were processed in the same batch for library preparation, sequencing, and analysis to control for batch effects. For each sample, 3 cDNA libraries for sequencing were generated per sample using the Collibri 3’ mRNA Library Prep Kit for Illumina (ThermoFisher). Libraries were quantified using the Collibri Library Quantification Kit (ThermoFisher) and evaluated on the Agilent 2100 Bioanalyzer using the Agilent High Sensitivity DNA Kit (Agilent Technologies). Libraries were sequenced on the Illumina NextSeq 500 using the NextSeq 500/550 High Output Kit (75 Cycles). Reads were filtered and trimmed using Cutadapt v3.4⁷⁶ and aligned with STAR v2.7.8a⁷⁷ to both hg38 (human reference genome) and mm39 (mouse reference genome). Aligned reads were sorted and indexed using SAMtools v1.12⁷⁸ and filtered to detect only human-specific reads using XenofilteR⁷⁹. Reads aligning to expressed genes were counted using HTSeq v0.13.5⁸⁰. Counts were CPM normalized and scaled per gene, prior to plotting. Heatmaps were generated using ggplot2v3.5.0.

Single-cell RNA sequencing and analysis

Fetal hematopoietic (CD45+) single-cell data from Bian et al.¹⁷ was downloaded from NCBI Gene Expression Omnibus, accession number GSE133345. For downstream analysis, primitive macrophage subpopulations (Mac 1, Mac 2, and Mac 3) were grouped together. UMAP plots were generated and colored by origin, Carnegie stage, and cell type grouping of all cells using Scanpy⁸¹.

Single-cell RNA sequencing was performed on MG01 cells (Line 6) (n = 3 independent biological replicates) and hiPSCs using 10X Genomics. The raw data for each dataset was the “Filtered Feature Barcode Matrix” output by Cell Ranger⁸² alignment and deconvolution pipeline. Each dataset was imported and mitochondrial content, number of genes by counts, and total counts were annotated for each cell. A knee locator-based method⁸³ was used to remove cells that have an overrepresentation of mitochondrial counts. The data was then filtered to include only cells that have a minimum of 1000 non-zero transcripts. Double detection and removal were performed and the dataset subset was to only include 5000 cells (so that each dataset has the same number of cells in the analysis). Each dataset was iteratively joined to the other datasets included in the analysis. Joining is performed on Ensembl IDs on an inner-join basis (only genes common to both datasets are included in the resulting data object). MG01 and hiPSC data were then joined to Bien et al. data on Ensembl IDs. Finally, the total counts per cell was normalized to 1,000,000 (CPM normalization), and log2 transformed.

PCA was run on the concatenated dataset using highly variable genes. Differential gene expression was run using Scanpy’s rank_genes_groups function and filtered based on a minimum log fold change = 4 and a threshold fraction of cells expressing the gene within and outside of the group = 0.5. Genes enriched in the Mac 1-3 groups and Mac 4 group were used for sparsity-based scoring. To score, the normalized expression matrix was binarized and the sparsity of the matrix was calculated. The matrix was subsetted to include the genes in the group of interest (Mac 1-3 or Mac 4). The gene set score for each cell was calculated as the sum of the binarized expression data for the gene group multiplied by the sparsity, normalized by the maximum score so that the value ranged from 0 to 1.

Differential gene expression was run on the Bian et al. dataset¹⁷ (with the same settings described previously) based on the cell type groupings of all cells to find unique markers for each group. Expression data from this dataset was standardized between 0 and 1 and shown as a heatmap. Unique genes found in the Mac 4 group and Mac 1-3 groups were displayed as dotplots using the single-cell RNA sequencing datasets and as a heatmap using engrafted MG01 bulk RNAseq expression data.

Software information

Images for brain sections acquired with Axioscan utilizing Zen 3.1. GAG levels, enzymatic activity, and LAMP1/2 assays were measured in the ClarioSTAR platform using MARS software v3.31. FACS data obtained in CytoFLEX LX using CytExpert v2.5 – only flow cytometry analysis performed (no cell sorting occured). Phagocytosis data was acquired and analyzed using the Incucyte platform v2023.1.1. Bulk RNAseq reads were filtered and trimmed using Cutadapt v3.4 and aligned with STAR v2.7.8a to both hg38 and mm39. Aligned reads were sorted and indexed using SAMtools v1.12 and filtered to detect only human-specific reads using XenofilteR (v1.6). Reads aligning to expressed genes were counted using HTSeq v0.13.5. scRNAseq raw data for each dataset was the “Filtered Feature Barcode Matrix” output by Cell Ranger v3.0.2 alignment and deconvolution pipeline UMAP plots were generated colored by origin, Carnegie stage, and cell type grouping of all cells using Scanpy 1.8.1. Biochemical and behavior graphs and statistical tests were performed in Prism v10.1.0. Pixel classification and cell counts performed with Zen 3.3 with Intellesis module. LAMP1 intensity per cell was measured using QuPath 0.4.3. Cell viability was measured with the Nucleocounter platform using Nucleoview v1.4.3

Miscellaneous software

All figures were created and assembled in Affinity Designer 1.10. Graphs and statistical tests performed in Prism v10.1.0.

Material availability

Cell lines #82 and #83 are not available for sharing because they are proprietary and reserved for potential clinical use. All critical experiments have been conducted with cell line #6, which was obtained via NHCDR (https://stemcells.nindsgenetics.org/). We generated additional IDUA, GUSB, NAGLU, and SGSH knockout iPSC cell lines based on parental cell line #6 which are available, subject to prior written authorization from the NHCDR and a material transfer agreement.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41467-024-52400-8