Study design
Sample size in experiments with mice was chosen according to previous experience with experimental models and assays. No sample or animal was excluded from the analyses. Mice were randomly assigned to each experimental group. Investigators were not blinded.
Plasmid construction
Plasmid containing LV transfer constructs were generated by standard cloning techniques starting from already available plasmids. LV.ET.mCherry was generated starting from LV.ET.GFP plasmid 35, digesting it with NheI-HF and BamHI-HF restriction enzymes (New England Biolabs) and inserting mCherry sequence obtained from a previously described plasmid36. Similarly, LV.ET.BFP was cloned by digesting LV.ET.GFP with SalI-HF and PmeI enzymes (New England Biolabs) and inserting BFP sequence obtained from a previously described plasmid36. LV.ET.CreERT2 was cloned by digestion of LV.ET.GFP plasmid using SalI-HF and NheI-HF enzymes and inserting CreERT2 seqeunces obtained by digestion of pMuLE ENTR SV40 CreERT2 L3-L2 plasmid (Addgene #62174) with SpeI and PspXI enzymes (New England Biolabs). LV.HCR-hAAT.GFP was obtained by exchanging the EcoRV-PGK-BamHI promoter from LV.PGK.GFP37 with the PCR-amplified EcoRV-HCR-hAAT-BglII from 767-RV-hAAT-wpre and 540B-ApoE-hAAT-LTR and pBSHCRHP-FIXIA, respectively38. The plasmid LV.HCR-hAAT.hFIX was obtained by first exchanging the XhoI-ET-NheI promoter with the PCR-amplified XhoI-HCR-hAATp-NheI into LV.ET.GFP and subsequently exchanging the NheI-GFP.wpre-SacII fragment from LV.HCR-hAAT.GFP with the XbaI-hFIX.wpre-SacII fragment from LV.CMV.hFIX plasmid39.
Cell culture
Human embryonic kidney (HEK) 293T cells were used for vector production and titration. Cells were maintained in culture with Iscove’s modified Dulbecco’s medium (IMDM, Corning) supplemented with 10% fetal bovine serum (FBS, Euroclone) and penicillin/streptomycin 100 international units (IU)/ml (Lonza). Cells were detached using PBS 0.05% trypsin 4mM EDTA solution. Cells were kept in a humidified incubator at 37°C 5% CO2.
LV production
Third-generation self-inactivating (SIN) LV were produced by calcium phosphate transient transfection of HEK 293T cells with a mix of transfer plasmid and the packaging plasmids pMDLg/pRRE, pCMV.REV, PMD2-VSV.G or pBA-AcMNPV-gp6440 and pAdVantage plasmids as previously described18. Culture medium was changed 14-16 hours after transfection and then collected after additional 30 hours. Supernatant was filtered with a 0.22 μm filter (Millipore) and ultracentrifuged in polyallomer tubes (Beckman Optima XL-100K Ultracentrifuge). In some cases, before centrifugation, LV-containing 0.22µm-filtered supernatants were pre-concentrated 4 times with Vivaflow 200 Tangential Flow Filtration Cassette 100 kDa (Sartorius), according to manufacturing instructions. LV-containing pellet was resuspended in the appropriate amount of phosphate saline buffer (PBS, Corning) to reach a 500X concentration, aliquoted and stored at -80°C. For IDLV production the same procedure was followed, but pMDLg/pRRE plasmid is substituted with pMDLg/p.RRE.D64Vint plasmid41
LV titration
LV infectious titer was calculated by in vitro transduction of 293T cells and quantitative polymerase chain reaction (qPCR). 105 293T cells were transduced with 10-fold serial dilutions of LV prepared in fresh medium in presence of polybrene (final concentration 8 μg/ml). 10 days after transduction cells were collected for VCN determination.
ddPCR for LV titration and VCN determination
Genomic DNA (gDNA) was extracted from transduced cells or organ sample using Maxwell 16 DNA purification Kit (Promega) following manufacturer’s instructions. gDNA from primary sorted cells was extracted using QIAamp DNA mini kit or QIAamp DNA micro kit (Qiagen), following manufacturer’s protocol. VCN per diploid genome was determined by analyzing 15-20 ng of gDNA through digital droplet PCR (ddPCR) with QX200 Droplet Digital PCR System (Biorad), according to manufacturer’s recommendations (each primer 900nM, probe 250nM). For quantification of LV genome copies the following primers and a probe were used:
HIV Fw: 5’-TACTGACGCTCTCGCACC-3’
HIV Rv: 5’- TCTCGACGCAGGACTCG-3’
HIV Pr 5’-(FAM)-ATCTCTCTCCTTCTAGCCTC-(MGBNFQ)-3’).
For quantification of human gDNA were used primers and probe designed on GAPDH gene (Applied Biosystems Hs00483111_cn) or the human telomerase gene:
Telo fw: 5’-GGCACACGTGGCTTTTCG-3’;
Telo rv: 5’-GGTGAACCTCGTAAGTTTATGCAA-3’;
Telo probe: VIC 5’-TCAGGACGTCGAGTGGACACGGTG-3’ TAMRA.
For quantification of mouse gDNA were used primers and probe designed on Sema3a gene:
Sema Fw: 5’- ACCGATTCCAGATGATTGGC-3’
Sema Rv: 5’- TCCATATTAATGCAGTGCTTGC-3’
Sema Pr: 5’-(HEX)-AGAGGCCTGTCCTGCAGCTCATGG-(BHQ1)-3’).
To calculate VCN, was used the formula VCN=(ng LV)/(ng gDNA)×2. Infectious titer is calculated with the formula TU⁄ml=(VCN×100.000)/(1⁄(dilution factor)).
Production, purification, and characterization of the AAV vectors
The AAV vectors used in this study are based on AAV type 2 backbone, and infectious vectors were prepared by the AAV Vector Unit at ICGEB Trieste (https://www.icgeb.org/avu-core-facility.html) in 293 cells by a cross-packing approach whereby the vector was packaged into AAV capsid 8, as described42. Viral stocks were obtained by CsCl gradient centrifugation. Titers were determined by measuring the copy number of viral genomes in pooled, dialyzed gradient fractions, as previously described42.
Mice experiments
C57BL/6 and NSG mice were purchased from Charles River Laboratories. Founder B6.129P2-Gt(ROSA)26Sortm1(CAG-Brainbow2.1)Cle/J mice28 (referred to as Confetti mice) were obtained from The Jackson Laboratories (stock #017492;). Founder SA+/CreERT2 mice (referred to as Alb-CreERT mice) were kindly provided by Daniel Metzger43. KC-DTR mice (referred to Clec4f-DTR mice) were kindly provided by Martin Guiliams27. All mice were maintained in specific pathogen-free conditions. Alb-CreERT/Confetti mice were obtained by crossing Confetti homozygous mice with Alb-CreERT homozygous mice. For experiments in Confetti mice or AlbCreERT2-Confetti mice, male and female were both used randomly, according to availability. For experiments in newborns and 2-week old C56BL/6 wild-type mice, male and female were used according to availability at birth, while for adult C57BL/6 mice only females were used. Blood samples were collected with capillary tubes from the retro-orbital plexus and supplemented with 0.38% sodium citrate buffer pH7.4 to avoid coagulation. LV administration in adult mice (6-10 weeks old) was carried out by tail-vein injection, in 2-weeks old mice by tail-vein or retro-orbital injection, in newborn mice (1-2 days old) by temporal vein injection. AAV administration in newborn mice was carried out by temporal vein injection. Tamoxifen (Sigma-Aldrich T5648) was dissolved in corn oil 10 mg/ml, stored at -20°C and administered subcutaneously in newborn mice and intraperitoneally in older mice at the dose of 0.1 mg/g. Diphtheria toxin (Millipore 322326) was resuspended in PBS 1μg/μl, stored at -20°C and administered intraperitoneally at a dose of 25ng/mouse diluted to 0.125ng/μl in PBS. Bortezomib (Velcade, 2.5mg/ml) was diluted 10X in saline solution NaCl 0.9% and administered intravenously at a dose of 1µg/g. EdU (Invitrogen A10044) was suspended in PBS at the final concentration of 10mM, stored at -20°C and administered subcutaneously in newborn mice and intraperitoneally in older mice. Mice were euthanized by cervical dislocation or CO2 inhalation. All protocols were approved by the Institutional Animal Care and Use Committee.
Fractionation and sorting of liver cell sub-populations
Fractionation of liver cell subpopulation was performed has previously described18. Liver was digested by perfusion via the inferior vena cava first with PBS 0.5 mM EDTA (~10 ml), then HBSS (Hank’s balanced salt solution, Gibco) 1% HEPES ((4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (~10 ml) and finally HBSS 1% HEPES 0.03% collagenase IV (Sigma) (~16.5 ml). Liver was then collected and passed through a 100 μm cell strainer (BD Biosciences). Cell suspension was centrifuged three times at 30, 25 and 20 g for 3 minutes to separate non-parenchymal cells (in the supernatant) from hepatocytes (in the pellet). Supernatant was centrifuged 7 min 650g and cells were loaded onto a 30%-60% Percoll (Sigma) gradient and centrifuged 1800g for 15 min. nPC interface was collected and washed twice with HBSS 1% HEPES. Hepatocyte fraction was stained with the following mix of antibodies, in 200 μl of volume:
Antigen
|
Fluorochrome
|
Clone
|
Company (code)
|
Dilution
|
CD31
|
APC
|
MEC 13.3
|
BD Biosciences (551262)
|
1:66
|
CD45
|
APC
|
30-F11
|
BD Biosciences (559864)
|
1:66
|
FC Block
|
|
2.4G2
|
BD Biosciences (553142)
|
1:40
|
Hepatocyte fraction was sorted to remove contaminants using BD FACSAria™ Fusion Cell Sorter (BD Biosciences) or MoFlo Astrios EQ Cell Sorter (Beckman Coulter). nPC fraction was first incubated with FC Block and then stained with the following mix of antibodies in 200 μl of volume:
Antigen
|
Fluorochrome
|
Clone
|
Company (code)
|
Dilution
|
CD31
|
FITC
|
MEC 13.3
|
BD Biosciences (551262)
|
1:30
|
CD45
|
e-fluor 450
|
30-F11
|
Invitrogen (48-0451-82)
|
1:30
|
F4/80
|
PE
|
CI:A3-1
|
BD Biosciences (553142)
|
1:30
|
CD45R/B220
|
PE-Cy5
|
RA3-6B2
|
BD Biosciences (553091)
|
1:30
|
CD11c
|
PE-Cy7
|
N418
|
Invitrogen (25-0114-82)
|
1:30
|
Enzyme-linked immunosorbent assay (ELISA)
Mouse blood was centrifuged 6500 rpm for 6 min 30 sec to collect plasma, which was stored at -80°C. Samples were diluted according to necessity. ELISA specific for human FIX antigen was used to determine its concentration in mouse plasma samples (Asserachrom IX:Ag, Stago), following manufacturer’s protocol. For IgG and IgM ELISA, plates were coated plates with 100ng of VSV.G (Vinci-Biochem V3-VSIG15-R-10) protein and of IgG and IgM antibodies for the standard curve and incubated 16 hours at 4°C. Plates were washed 3 times with PBS 0.05% Tween 20 and incubated for 2 hours at room temperature with 200µl of blocking solution (PBS 0.05% Tween-20 1% Chicken Ova). Plates were washed again 3 times and the incubated with 100 µl of diluted samples and incubated 2 hours at 37°C. Plates were washed 4 times and anti-IgG (1:10000; Sigma A2554-1ML) or anti-IgM (1:5000, Southern Biotech 1140-05), HRP conjugated detection antibodies were added and incubated 1 hour at 37°C. Plates were washed 4 times and incubated with 100 µl of TMB for 10 minutes and stopped with 50 µl of HCl. Plates were red at 450nm using a Multiskan GO microplate reader (Thermo Fisher Scientific).
Luciferase assay
For luciferase assay Pierce™ Gaussia Luciferase Flash Assay Kit (Thermo Scientific 16159) was used according to manufacturer’s protocol. Plates were red using FLUOstar Omega plate reader (BMG Labtech).
Immunofluorescence imaging
Livers were harvested from mice, washed briefly in PBS and fixed 4 hours in PBS 4% paraformaldehyde (PFA), then washed again briefly in PBS before being stored at least 24 hours in 30% sucrose 0.02% sodium azide in H2O. Livers were the frozen in OCT (optimal cutting temperature) compound (Killik, Bio Optica) and slices 5 to 20 μm thick were cut at cryostat (Histo-line MC5050), placed on Superfrost® Plus microscope slides (Thermo Scientific) and stored at -80°C. All the staining steps were performed protected from lights to preserve endogenous fluorescence. Slides were thawed at room temperature for at least 2 hours, then washed 3 times in PBS 0.1% X-Triton. Edges were drawn around the tissue using immunostaining pap pen (Sigma-Aldrich) to contain staining solutions. Blocking step was performed using PBS 0.1% X-Triton 1% bovine serum albumin (BSA) 5% FBS for 1 hour at room temperature in a humid chamber. Blocking solution was then substitute with mix containing primary antibodies in blocking solution and incubation was carried on for 12-16 hours at 4°C in a humid chamber. Primary antibodies mix solution was removed and slides were washed 3 times in PBS 0.1% X-Triton. Secondary antibodies were diluted in blocking solution together with Hoechst and incubated for 1 hour at room temperature in a humid chamber. Slides were then washed in PBS and coverslip was added using Fluoromount-G (Invitrogen) as mounting medium. Slides were dried for 16 hours at room temperature or for longer at 4°C and stored at 4°C before acquisition. EdU staining was performed using Click-iT™ EdU Cell Proliferation Kit with AlexaFluor 488 dye (Invitrogen C10337) according to manufacturer’s protocol. Antigen unmasking was performed by incubating slides 20 min in citrate buffer (10mM Citric Acid, pH 6.0) at 98°C. For nuclear staining concentration of X-Triton has been raised to 0.5% for blocking buffer and washing solution. The following antibodies were used for immunofluorescence staining:
Antigen
|
Species
|
Fluorochrome
|
Clone
|
Company (code)
|
Dilution
|
GFP
|
Rabbit
|
-
|
Polyclonal
|
Invitrogen (A11122)
|
1:1000
|
GFP
|
Chicken
|
-
|
Polyclonal
|
Invitrogen (A10262)
|
1:250
|
mCherry
|
Rabbit
|
-
|
Polyclonal
|
Abcam (ab167453)
|
1:200
|
mCherry
|
Chicken
|
-
|
Polyclonal
|
Abcam (ab205402)
|
1:1000
|
CK7
|
Rabbit
|
-
|
EPR17078
|
Abcam (ab181598)
|
1:250
|
GS
|
Rabbit
|
-
|
Polyclonal
|
Novus Bio. (NB110-4104)
|
1:5000
|
F4/80
|
Rat
|
-
|
CI:A3-1
|
Abcam (ab6640)
|
1:200
|
HNF4α
|
Rabbit
|
-
|
C11F12
|
Cell Sig. (3113S)
|
1:50
|
E-cadherin
|
Goat
|
-
|
Polyclonal
|
R&D systems (AF748)
|
1:50
|
Rabbit IgG
|
Goat
|
AF488
|
Polyclonal
|
Invitrogen (A11034)
|
1:1000
|
Rabbit IgG
|
Goat
|
AF546
|
Polyclonal
|
Invitrogen (A11010)
|
1:1000
|
Rabbit IgG
|
Donkey
|
AF647
|
Polyclonal
|
Invitrogen (A31573)
|
1:1000
|
Chicken IgY
|
Goat
|
AF488
|
Polyclonal
|
Invitrogen (A11039)
|
1:1000
|
Chicken IgY
|
Goat
|
AF561
|
Polyclonal
|
Invitrogen (A11040)
|
1:1000
|
Rat IgG
|
Goat
|
AF561
|
Polyclonal
|
Invitrogen (A11081)
|
1:1000
|
Rat IgG
|
Goat
|
AF647
|
Polyclonal
|
Invitrogen (A21247)
|
1:1000
|
Hoechst
|
-
|
-
|
-
|
Invitrogen (H3570)
|
1:2000
|
Images were acquired using confocal microscope Leica TCS SP5 (LAS AF software), Leica TCS SP8 (LAS X software) or Mavig Rs-G4 (RSG4 software) at 20X or 40X magnification. Images were analyzed using Image J or MATLAB software. Identification of stained cells was performed by manually selecting a threshold and counting objects bigger than 10 μm2 for nuclei and 100 μm2 for hepatocytes. Total liver area was measured by saturating signal from the whole tissue. Cells positive for two or more colors were identified by calculating pixels positive in two or more channels using Image J. Cluster dimension was measured using Image J, the number of cells per cluster was calculated by dividing the dimension of each cluster by the average size of a hepatocyte at each age, that was calculated by segmenting clusters in representative images and measuring average cell dimension (D3=200 μm2, W1=250 μm2, W2=300 μm2, W3=350 μm2, W4, W6, Y1=450 μm2). Peri-portal and peri-central areas were identified by marking tissue area within a fixed distance from CK7+ or GS+ area using MATLAB: CK7, GS areas and hepatocytes where segmented form their respective fluorescent images by first filtering out background fluorescence and small-scale noise and then applying a threshold (using Image J). An additional filter on the dimension ensures to discard unreasonable small or large objects (<100 µm2, outliers area distribution). The whole tissue is also segmented through a very low threshold detecting auto fluorescence; unwanted areas like fringed borders are manually excluded. PP and PC regions are defined as morphological dilations of CK and GS segmented areas of the specified radii (D1 187 µm, W2 200 µm, W8 225 µm) that follow within the tissue. We finally measure the fraction of hepatocyte’s area in the PP and PC regions.
Immunohistochemistry imaging
Livers were perfused from the inferior vena cava with PBS EDTA 0.5mM to remove blood, then harvested and fixed at least 24h in zinc-formalin. Embedding and cutting of livers and slides preparation, staining and acquisition were performed by Centro di Imaging Sperimentale facility in San Raffaele hospital. Briefly, formalin-fixed paraffin-embedded consecutive sections (4 um) were dewaxed and hydrate through graded decrease alcohol series and stained for histology or immunohistochemical characterization (IHC). Slides were immunostained with Automatic Leica BOND RX system (Leica Microsystems GmbH, Wetzlar, Germany). First, tissues were deparaffinized and pre-treated with the Epitope Retrieval Solution (ER1 Citrate Buffer). Anti-GFP primary antibody (Invitrogen, A11122) was incubated 30 minutes at room temperature (1:1000) and was developed with Bond Polymer Refine Detection (Leica, DS9800). Slides were acquired with Aperio AT2 digital scanner at magnification of 20X (Leica Biosystems) and analyze with Imagescope (Leica Biosystem).
Liver clearing and 3D imaging
Livers were perfused with PBS 5mM EDTA (~10 ml) and PBS 4% PFA (150 ml), were then harvested and incubate in PBS 4% PFA for 30 min and then moved in PBS and stored at room temperature protected from light. X- CLARITY™ system (Logos Biosystems) was used for liver clarification44. Liver lobes were separated and incubated in X-CLARITY™ Hydrogel Solution supplemented with X-CLARITY™ Polymerization Initiator for 24 hours at 4°C protected from light in a 6-weel plate. The plate was then moved into the X-CLARITY™ Polymerization System and incubated for 3 hours at 37°C in vacuum conditions (-90kPa). Liver lobes were stored into Electrophoretic Tissue Clearing Solution until the next step was performed. Liver lobes were then cut into smaller pieces, placed in holder chamber and moved into X-CLARITY™ Tissue Clearing System II for electrophoresis (4.5h, 1.4 A, 37°C). Samples were stored in PBS at room temperature protected from light. 2 hours before acquisition samples were moved into X-CLARITY™ Mounting Solution. Acquisition was performed using Nikon A1 MP 2-photon microscope or Zeiss Lightsheet Z.1.
Electron microscopy analysis
Mice livers were perfused via inferior vena cava with PBS-EDTA 0.5mM and then with fixation buffer (2.5% glutaraldehyde, 2% paraformaldehyde, 2mM CaCl2, 2% sucrose in 0.1M cacodylate buffer). Livers were excised and further fixed by immersion in the same fixation buffer overnight. Fixed liver lobules were sectioned into 150µm thick sections using a vibratome (Leica Vibratome VT1000s). Sections were further post-fixed in 1% osmium tetroxide in 0.1 M Na cacodylate buffer, dehydrated through a series of increasing concentration of ethanol, immersed in absolute hexamethyldisilazane (HMDS) that was left to evaporate overnight. Dried sections were sputter coated with chromium using a Quorum Q150T ES sputter coater. Sections were then mounted on scanning electron microscope (SEM) stubs using conductive adhesive tape and observed in a field-emission SEM Gemini 500, (ZEISS, Oberkochen, Germany) at University of Milan Bicocca or University of Milan.
Integration site analysis
The PCR procedure used to obtain and map integration sites (IS) is consistent with the method previously described45. Briefly, Sonication Linker Mediated (SLiM)-PCR was performed on DNA extracted from murine liver. The DNA was sheared using a Covaris E220 Ultrasonicator, generating fragments with an average size of 1000bp. Each sonication product was split in 3 technical replicates. The fragmented DNA was subjected to end repair and 3’ adenylation, and then ligated to linker cassettes containing a 8 nucleotide sequence barcode used for sample identification, a random 12 nucleotides sequence needed for clonal abundance estimation, and all the sequences required for the Read 2 Illumina paired end sequencing. The ligation product is then subjected to 35 cycles of exponential PCR using primers specific for the lentiviral vector LTR and the Linker cassette. The amplification product is then re-amplified with additional 10 PCR cycles using a primer specific for the linker cassette and the LTR. These primers contain a 8-nucleotide barcode used for sample identification (coupled with the barcode on the linker cassette) and the sequences needed for the Read 1 Illumina sequencing, plus a 12 random nucleotides sequence allowing easier cluster recognition in the first sequencing cycles on the Illumina sequencer. The generated SLiM-PCR products are thus associated to a unique pair of barcodes, assembled into libraries subjected to Illumina sequencing. Sequencing reads were processed using a custom bioinformatics pipeline (VISPA2)46 that isolates genomic sequences flanking the vector LTR and maps them to the murine genome (mm9). Clonal abundance estimates as the relative percentage of genome numbers was determined by the R package "sonicLength”47; Downstream analyses of vector IS, such as Shannon diversity index, relative abundance analysis and common IS (CIS) analysis, were performed using ISAnalytics48. CIS were calculated by the Grubbs test for outliers49.
RNA sequencing
For bulk transcriptomic analysis RNA was extracted using RNeasy Micro Kit (Qiagen), following manufacturer’s instructions. Library preparation and Illumina sequencing was performed at Azenta Lifesciences. RNA-seq reads were analyzed with FastQC (v0.11.6) to assess the quality and pre-processed using Cutadapt (v0.6.6) to get rid of low-quality bases. Resulting reads were then aligned to the mouse reference assembly (GRCm39) using the STAR software (v2.7.6a) with standard parameters, exploiting the Genocde annotation (vM29). Gene abundancies were calculated using the Subread featureCounts function (v2.0.1), and Differential Gene Expression (DGE) analysis comparing mCherry-positive vs. -negative samples was performed using the R/Bioconductor package DESeq2 (v1.30.0), normalizing for library size using DESeq2’s median of ratios. p-values were corrected using false discovery rate (FDR). Genes having FDR < 0.01 and |logFC| > 1 were considered as differentially expressed. Post-analyses on DGE results were performed with the R/Bioconductor package ClusterProfiler (v4.7.1), using both Panglao and GO Biological Processes as reference databases.
Spatial transcriptomic
Livers were collected and snap frozen in isopentane, then embedded in Killik OCT on dry ice. 5-µm slices were cut using Leica cryostat CM1860 and put on commercial Visium 10X slides. Tissue permeabilization optimization (Visium Spatial Tissue Optimization Slide & Reagents Kit, 1000193) and library preparation (Visium Spatial Gene Expression Slide & Reagents Kit, 1000187) were performed according to manufacturer’s protocol. Illumina sequencing was performed at the San Raffaele Center for Omics Sciences. Input samples were analyzed with the Space Ranger software (v1.2.2) from 10X Genomics to obtain the corresponding input FASTQ files (using the mkfastq utility), and then the spot-by-gene matrices (using the count utility). These latter outputs were produced by aligning input (exploiting STAR) and then quantifying the barcode/UMI per sport, which were imported (with the corresponding tissue slide image) and analyzed with Seurat (v4.1.1). In details, samples were merged by age (i.e., newborns, 2-week-old, and adults) and all together, and each combined object was analyzed by performing normalization, principal component analysis, batch removal (using Harmony), clustering and UMAP embedding computation. Cluster makers and differential gene expression among ages were computed using the Seurat functions FindAllMarkers and FindMarkers, respectively. Enrichment analysis on (positive) differentially expressed genes resulting from the different comparisons was performed with the R/Bioconductor package ClusterProfiler (v4.7.1), using both Panglao and KEGG as reference databases. Information on liver layers was integrated by using as reference a previously created object on Halpern et al. data to find common anchors with each sample dataset (using the Seurat function FindTransferAnchors), and then mapping the labels corresponding to layers on each spot (using the Seurat function TransferData). Pseudotime analysis on the full combined object was performed with Monocle3 using as starting point the layer belonging to newborn samples to infer the overall trajectory.
Statistical analysis
Statistical analyses were performed upon consulting with professional statisticians at the San Raffaele University Center for Statistics in the Biomedical Sciences (CUSSB). When normality assumptions were not met, non-parametric statistical tests were performed. Two-tailed Mann- Whitney test was performed to compare two independent groups, while in presence of more than two independent groups Kruskal-Wallis test followed by post-hoc analysis (Dunn’s test for multiple comparisons against the reference control group along with Bonferroni’s correction) was applied. For repeated measures over time, two-way ANOVA was performed. For paired observations, Wilcoxon matched-pairs signed rank test or Friedman test with Dunn’s multiple comparisons test were performed. To compare a group to a theoretical value one-sample Wilcoxon signed-rank test was applied. Inferential techniques were applied in presence of adequate sample sizes (n ≥ 5), otherwise only descriptive statistics are reported. Growth curves and repeated measures data have been modelled using linear mixed-effects models (LME, Pinheiro & Bates, 2000), including random-effect terms for capturing underlying heterogeneity associated with both experimental observations and experimental session (non-nested or nested random effects). Outcome variables have been transformed using standard transformations (logarithm, square/cubic root, ordered quantile normalization) in order to satisfy LME assumptions. Post-hoc analysis after LME was performed to compare in a pairwise fashion treatment groups at a fixed time point. Bonferroni’s approach was used to adjust p-values for multiplicity. For all analyses the significance threshold was set at 0.05. Analyses were performed using GraphPad Prism and R statistical software (version 4.1.2, https://cran.r-project.org/index.html).