Triplication of the interferon receptor locus contributes to hallmarks of Down syndrome in a mouse model

Abstract

Down syndrome (DS), the genetic condition caused by trisomy 21, is characterized by variable cognitive impairment, immune dysregulation, dysmorphogenesis and increased prevalence of diverse co-occurring conditions. The mechanisms by which trisomy 21 causes these effects remain largely unknown. We demonstrate that triplication of the interferon receptor (IFNR) gene cluster on chromosome 21 is necessary for multiple phenotypes in a mouse model of DS. Whole-blood transcriptome analysis demonstrated that IFNR overexpression associates with chronic interferon hyperactivity and inflammation in people with DS. To define the contribution of this locus to DS phenotypes, we used genome editing to correct its copy number in a mouse model of DS, which normalized antiviral responses, prevented heart malformations, ameliorated developmental delays, improved cognition and attenuated craniofacial anomalies. Triplication of the Ifnr locus modulates hallmarks of DS in mice, suggesting that trisomy 21 elicits an interferonopathy potentially amenable to therapeutic intervention.

Fig. 1. Overexpression of IFNRs associates with inflammatory signatures in DS.

a, Manhattan plots of human chromosomes 3 and 21 (HSA3 and HSA21) displaying results of whole-blood transcriptome analysis for individuals with trisomy 21 (T21, n = 304, 163 male and 141 female) versus euploid controls (D21, n = 96, 44 male and 52 female). Red points mark DEGs identified by DESeq2. b, Sina plots displaying results for representative DEGs. Boxes represent interquartile ranges and medians, with notches approximating 95% CIs; q values determined by DESeq2 with Benjamini–Hochberg correction. c, Heatmaps displaying top left—NES from GSEA of transcriptome changes in individuals with DS. Only the top ten positively enriched pathways by NES are shown; top right—NES from GSEA of transcriptome signatures associated with expression of HSA21 genes surrounding the IFNR cluster (red). Spearman correlations were defined for HSA21-encoded mRNAs versus all other mRNAs using only trisomy 21 samples and ρ values as the GSEA ranking metric; middle—Spearman correlations between mRNAs encoded on HSA21 versus mRNAs for indicated DEGs encoded elsewhere in the genome among individuals with DS; bottom—Spearman correlations between mRNAs encoded on HSA21 and plasma levels of CRP and IL6 in individuals with DS. The asterisks indicate q < 0.1 from either GSEA or Spearman correlations with permutation test and Benjamini–Hochberg correction. d,e, Volcano plots of Spearman correlations for (d) IFNAR1 or (e) DYRK1A mRNA abundance versus all other mRNAs among individuals with DS. Heatmaps display NES from GSEA of ranked Spearman ρ values. f, Scatter plots displaying relationships between expression of IFNAR1 versus indicated DEGs among individuals with DS. g, Volcano plot of Spearman correlations for CRP levels versus mRNAs encoded on HSA21 among individuals with DS (n = 249, 137 male and 112 female). h, Scatter plots displaying relationships between CRP and two example mRNAs encoded on HSA21. In f and h, individual points are colored by local density and blue lines represent linear regression fits with 95% CIs in gray; ρ and q values from Spearman correlation with permutation test and Benjamini–Hochberg correction. RPKM, reads per kilobase per million; NES, normalized enrichment scores; CI, confidence interval.

Fig. 2. Triplication of the Ifnr locus contributes to global dysregulation of gene expression in a mouse model of DS.

a, Top—diagram indicating genomic locations of the mouse Ifnr gene cluster on MMU16 and gRNAs (orange arrowheads) employed for genome editing using CRIPSR–Cas9 technology. Positions are indicated in base pairs (bp) for the GRCm38 assembly of the M. musculus genome. Bottom—copy number variant analysis from WGS for a candidate founder (F0) bearing a deletion relative to a WT control. b, Breeding strategy to correct copy number of the Ifnr gene cluster in the Dp16 mouse model of DS. c, Volcano plots showing transcriptome analysis of mesenteric lymph nodes obtained from naïve adult WT (n = 5, 2 male and 3 female), Dp16 (n = 6, 3 male and 3 female), and Dp16^2xIfnrs (n = 6, 3 male and 3 female), highlighting expression of Ifnrs (red), other MMU16 genes triplicated in Dp16 (blue), with DEGs encoded elsewhere in the genome in gray, and all other genes in black. d, Expression levels in RPKM for representative MMU16-encoded mRNAs from mesenteric lymph nodes. q Values defined by DEseq2 after Benjamini–Hochberg correction. e, Scatter plots comparing mRNA fold-changes for Dp16 DEGs in mesenteric lymph nodes (top, sample sizes as in c) and brain (bottom, WT (n = 6, 2 male and 4 female), Dp16 (n = 5, 2 male and 3 female), and Dp16^2xIfnrs (n = 7, 4 male and 3 female)) for Dp16/WT and Dp16^2xIfnrs/WT, with Ifnrs highlighted in red, Dp16 triplicated genes in blue, nontriplicated Dp16 DEGs in gray, and slope (m) colored accordingly. Solid gray lines represent linear fits for the nontriplicated Dp16 DEGs. f, Sina plots displaying absolute fold-changes for DEGs triplicated in Dp16 (blue, excluding the four Ifnrs), and nontriplicated Dp16 DEGs (gray) across the genome for Dp16 versus WT or Dp16^2xIfnrs, comparisons, with P values for two-sided paired Wilcoxon rank tests, boxes representing interquartile ranges and medians, and notches approximating 95% CIs. gRNA, guide RNA.

Fig. 3. Triplication of the Ifnr locus drives increased IFNR expression and exacerbated antiviral responses in a mouse model of DS.

a, gMFI relative to WT mice, as measured by flow cytometry, for IFNR proteins on CD45⁺ WBCs from heterozygous Ifnr knockout mice (WT^1xIfnrs, n = 5, 3 male and 2 female), WT (n = 8 for IFNAR1, 2 male and 6 female, n = 16 for IFNGR2, 10 male and 6 female), Dp16 (n = 7, 5 male and 2 female) and Dp16^2xIfnrs (n = 6 for IFNAR1, 4 male and 2 female; n = 8 for IFNGR2, 6 male and 2 female). Horizontal dashes indicate mean values. Significance determined by two-sided Mann–Whitney test. b, gMFI relative to WT, as measured by flow cytometry, for phosphorylated STAT1 in WBCs at baseline or after 30 min stimulation with IFN-α or IFN-γ. Number of animals—unstimulated WT (n = 25, 16 male and 9 female), Dp16 (n = 15, 7 male and 8 female) and Dp16^2xIfnrs (n = 8, 8 male); +IFN-α WT (n = 23, 14 male and 9 female), Dp16 (n = 12, 3 male and 9 female) and Dp16^2xIfnrs (n = 9, 9 male); +IFN-γ WT (n = 25, 16 male and 9 female), Dp16 (n = 12, 12 male) and Dp16^2xIfnrs (n = 7, 7 male). Significance determined by two-sided Mann–Whitney test. c, Heatmap displaying NES from GSEA with Hallmark gene sets of transcriptome fold-changes for the indicated comparisons in mesenteric lymph nodes, sorted by NES for Dp16/WT; *q < 0.1 defined by GSEA with Benjamini–Hochberg correction. d, Heatmaps (top) representing median RPKM expression z scores per genotype and sina plots (bottom) for example genes from the indicated gene sets; q values determined by DESeq2 with significance defined as q < 0.1 after Benjamini–Hochberg correction. e, Kaplan–Meier plot comparing survival across genotypes during chronic stimulation with the TLR3 agonist poly(I:C); significance determined by Mantel–Cox log-rank test. f, Percentage weight loss normalized to total number of poly(I:C) doses. For e and f, WT sham (n = 9, 2 male and 7 female), WT poly(I:C) (n = 13, 6 male and 7 female), Dp16 poly(I:C) (n = 9, 6 male and 3 female), Dp16^2xIfnrs poly(I:C) (n = 13, 6 male and 7 female). g, TNF protein in serum on day 3 of poly(I:C) exposure, WT sham (n = 6, 2 male, 4 female), WT poly(I:C) (n = 7, 3 male, 4 female), Dp16 poly(I:C) (n = 6, 5 male, 1 female) and Dp16^2xIfnrs poly(I:C) (n = 6, 3 male and 3 female). In f and g, horizontal dashes indicate group means and P values for pairwise comparisons were determined by two-sided Mann–Whitney test, with significance set at P < 0.05. h, Scatter plot comparing TNF concentration and percent weight loss on day 3 of poly(I:C) exposure (n = 25 animals, numbers by genotype and sex as in g), with simple linear regression fit line. ρ and P values from Spearman correlation with permutation test. gMFI, geometric mean fluorescent intensities. Source data

Fig. 4. Triplication of the Ifnr locus is necessary for increased incidence of heart malformations in a mouse model of DS.

a–c, Representative images of hematoxylin and eosin stained serial sections through entire mouse hearts at embryonic day (E)15.5, showing (a) normal septation of the four heart chambers in a WT embryo, (b) an ASD in a Dp16 embryo and (c) a VSD in a Dp16 embryo. A total of 58 formalin-fixed paraffin-embedded embryos were processed and analyzed across four independent batch experiments. d, Heart malformation frequencies at E15.5. Other—outflow tract anomaly. P values were calculated for differences in CHD occurrence for pairwise comparisons between genotypes using two-sided Fisher’s exact test. Number of animals—WT (n = 24, 12 male and 12 female), Dp16 (n = 18; 9 male, 7 female and 2 undetermined sex) and Dp16^2xIfnrs (n = 16, 4 male and 12 female). e, Scatter plots comparing mRNA fold-changes for Dp16 DEGs in heart tissue from mice at the indicated ages for Dp16/WT and Dp16^2xIfnrs/WT, with Ifnrs highlighted in red, Dp16 triplicated genes in blue, nontriplicated Dp16 DEGs in gray and slope (m) colored accordingly; solid gray lines represent linear fits for the nontriplicated Dp16 DEGs. Number of animals—E12.5 hearts WT (n = 6, 2 male and 4 female), Dp16 (n = 6, 3 male and 3 female) and Dp16^2xIfnrs (n = 6, 5 male and 1 female); E18.5 hearts WT (n = 6, 4 male and 2 female), Dp16 (n = 6, 3 male and 3 female) and adult hearts WT (n = 5, 2 male and 3 female), Dp16 (n = 6, 3 male and 3 female) and Dp16^2xIfnrs (n = 5, 3 male and 2 female). f, Heatmap of GSEA for transcriptome changes in E12.5 hearts, sorted by NES for Dp16 versus WT (sample sizes described in e); asterisks indicate q < 0.1 from GSEA. g, Heatmaps (top) representing median RPKM expression z scores per genotype and sina plots (bottom) for example genes from the indicated gene sets; q values determined by DESeq2. R, right; L, left; A, atrium; V, ventricle. Source data

Fig. 5. Triplication of the Ifnr locus promotes developmental delays and cognitive deficits in a mouse model of DS.

a, Odds ratio plots for developmental milestone achievement in neonates as assessed by mixed effects Cox regression for the indicated pairwise comparisons between Dp16, Dp16^2xIfnrs and WT animals, with adjustment for the covariates sex (fixed) and litter (random). Square points represent ‘success’ ratios with size proportional to −log₁₀(q) and error bars corresponding to 95% CIs; red indicates q < 0.1 after Benjamini–Hochberg correction; vertical dashed lines indicate odds ratio of 1. Numbers of animals assessed for each milestone are shown in the table at right. b, Freezing behavior during CFC of adult WT (n = 33, 13 male and 20 female), Dp16 (n = 17, 8 male and 9 female) and Dp16^2xIfnrs (n = 23, 11 male and 12 female). Data are represented as means ± s.e.m., with significance determined by two-way repeated measures ANOVA and Tukey’s HSD test; asterisks indicate P < 0.05 and are colored by comparison—Dp16 versus WT (blue), Dp16^2xIfnrs versus WT (orange) and Dp16 versus Dp16^2xIfnrs (pink). See Source Data and Extended Data Fig. 5 for exact P values. c, Swim path efficiency of mice navigating to the escape platform in an MWM for male and female (M and F), WT (n = 39, 19 male and 20 female), Dp16 (n = 28, 13 male and 15 female) and Dp16^2xIfnrs (n = 25, 12 male and 13 female). Statistics as in b. d, Representative swim trials for males from block 9 in the MWM with platform location denoted for the acquisition (gray circle) and reversal (black circle) phases. e, Mouse target quadrant occupancy during reversal probe trial of MWM; data are represented as means ± s.e.m., with significance determined by one-way ANOVA and Dunnett’s correction. f, Heatmap displaying P values from IPA of transcriptome changes in brains from Dp16 versus WT animals, ranked by decreasing significance (right-tailed Fisher’s exact test, WT (n = 6, 2 male and 4 female), Dp16 (n = 5, 2 male and 3 female) and Dp16^2xIfnrs (n = 7, 4 male and 3 female). g, Heatmaps (top) representing median RPKM expression z scores per genotype and sina plots (bottom) for the example genes from the indicated pathways; q values determined by DESeq2, with significance set at q < 0.1. Source data

Fig. 6. Triplication of the Ifnr locus exacerbates craniofacial anomalies in a mouse model of DS.

a, Representative lateral views of μCT reconstructions of skulls from WT, Dp16 and Dp16^2xIfnrs animals, aligned and scaled based on the same 3D linear measurement. b, Lateral (top) and dorsal (middle) views of the outer portion of the skull proper transparently overlaid on cranial base interior view from a WT mouse, with landmarks on the skull (top and middle) and mandible (bottom) in yellow, and interior landmarks on the cranial base in turquoise. Smaller and larger interlandmark distances (blue and green, respectively) in WT relative to Dp16 or Dp16^2xIfnrs calculated by EDMA. Number of animals—WT (n = 7, 2 male and 5 female), Dp16 (n = 7, 4 male and 3 female) and Dp16^2xIfnrs (n = 6, 4 male and 2 female), followed by bootstrapping 10,000 times. Turquoise—distances different in WT relative to Dp16 rescued in Dp16^2xIfnrs. Red—distances with no difference in WT relative to Dp16 but different in Dp16^2xIfnrs. c, Example interlandmark distances before bootstrapping. Number of animals as in b. P values determined by one-way ANOVA with Tukey’s HSD test, with significance set at P < 0.05; horizontal dashes indicate means. d, Form difference ratios of mean population estimates for distances on the skull after bootstrapping, with colors assigned to any distance that includes a cranial base landmark. e, Images of cranial base interior views for WT and Dp16 displaying normal or completely fused ISS. f, Frequency of complete ISS fusion (black) compared between cohorts by pairwise two-sided Fisher’s exact test with significance set at P < 0.05. Number of animals—WT (n = 20, 12 male and 8 female), Dp16 (n = 11, 6 male and 5 female) and Dp16^2xIfnrs (n = 14, 7 male and 7 female). g,h, Scatter plots comparing ISS fusion scores to BS length (g) or midface length (h) for each skull where total n = 20 (numbers per genotype and sex as in b), showing Spearman ρ and P values (permutation test). Black lines represent linear fits with 95% CIs shaded in gray. i, Heatmap displaying median RPKM expression of z scores for example genes dysregulated in E10.5 facial mesenchyme. Number of animals—WT (n = 3, 2 male and 1 female), Dp16 (n = 3, 2 male and 1 female) and Dp16^2xIfnrs (n = 3, 2 male and 1 female). j, Sina plots displaying expression levels for example genes; q values determined by DESeq2 with Benjamini–Hochberg correction, with significance set at q < 0.1. PS, presphenoid bone. Source data

Extended Data Fig. 1. Overexpression of IFNRs associates with inflammation in people with trisomy 21.

a, Schematic of biospecimen source and processing for datasets underlying Fig. 1 and Extended Data Fig. 1. b, Age distribution of participants by karyotype. Trisomy 21: T21, Down syndrome (n = 304, 163 male and 141 female). D21: euploid controls (n = 96, 44 male and 52 female). c, Top: Heatmap displaying the results of Gene Set Enrichment Analysis (GSEA) of ranked rho values from Spearman correlation analysis of mRNAs encoded on HSA21 versus all other mRNAs in the whole blood transcriptome of individuals with T21 (n = 304). NES: normalized enrichment score. Significance defined by GSEA as q < 0.1 after Benjamini-Hochberg correction. Middle: Spearman correlations between mRNAs encoded on HSA21 and the indicated differentially expressed genes encoded elsewhere in the genome among individuals with T21 (n = 304). Bottom: Spearman correlations between mRNAs encoded on HSA21 and the plasma levels of CRP and IL6 proteins in individuals with T21 (n = 249, 137 male and 112 female). q < 0.1 from Spearman with permutation test and Benjamini-Hochberg correction for middle and bottom. d, Distribution of NES values of GSEA run on Spearman ρ value matrices to assess correlations between expression of mRNAs encoded on HSA21 and top 10 gene sets elevated in the transcriptome of individuals with T21 (n = 304). Data are presented as sina plots where boxes represent interquartile ranges and medians, and notches approximate 95% confidence intervals. e, Correlations between indicated mRNAs among individuals with T21 (n = 304). f, Volcano plot of ρ and q-values from Spearman correlations of IL6 protein abundance in plasma versus expression of HSA21 genes in the transcriptome of individuals with T21 (left) and scatter plots showing Spearman correlations between the indicated mRNAs and IL6 (right) (n = 249). In e, f, statistical significance defined as q < 0.1 by Spearman correlations with permutation test and Benjamini-Hochberg correction. h. In all scatter plots, blue lines represent linear regression fits with 95% confidence intervals in gray. In the volcano plot in f, the triangle indicates q = 0.

Extended Data Fig. 2. Mouse model to determine if triplication of the Ifnr locus on mouse chromosome 16 is necessary for Down syndrome phenotypes.

a, Schematic of the mouse interferon receptor (Ifnr) gene locus and CRIPSR/Cas9 guide RNAs (gRNAs, orange arrowheads) used for deletion of 192 kb genomic locus on chromosome 16 (MMU16). Black arrows indicate forward (FW) and reverse (REV) primers. Location of genes based on GRCm38 reference genome is indicated on an ideogram of MMU16 cytogenetic regions colored according to Giemsa banding. b, Representative PCR from DNA of wild-type (WT) mouse, and mice heterozygous or homozygous for the expected knock-out (KO) deletion (WT^1xIfnrs or WT^0xIfnrs, respectively). Gel image represents example of genotyping PCR used to characterize >50 pups derived from modified zygotes. This approach screened 5113 different descendants to date from a single heterozygous founder (F0). c, Representative Sanger sequencing of the single modified allele transmitted from a F0 WT^1xIfnrs male to the first generation of progeny (F1) after inter-crossing with a WT female. d, Whole genome sequencing followed by copy number variant analysis of the F0 WT^1xIfnrs with site of deletion on MMU16 denoted by arrow (left) that is absent when two non-related C57BL/6 N WT mice are compared (right). Significance was determined by CNV-seq (*p < 0.1). e, Transcriptome analysis of hearts at embryonic day (E)12.5 - WT (n = 6, 2 male, 4 female), Dp16 (n = 6, 3 male, 3 female), Dp16^2xIfnrs (n = 6, 5 male, 1 female), E18.5 - WT (n = 6, 4 male, 2 female), Dp16 (n = 6, 3 male, 3 female), Dp16^2xIfnrs (n = 6, 3 male, 3 female), adult - WT (n = 5, 2 male, 3 female), Dp16 (n = 6, 3 male, 3 female), Dp16^2xIfnrs (n = 5, 3 male, 2 female), and facial mesenchyme at E10.5 - WT (n = 3, 2 male, 1 female), Dp16 (n = 3, 2 male, 1 female), Dp16^2xIfnrs (n = 3, 2 male, 1 female). Sina plots of mRNAs encoded on MMU16 triplicated in Dp16 excluding the four Ifnrs (top), and other differentially expressed genes (DEGs) across the genome for Dp16 versus WT (bottom). p-values calculated by two-sided paired Wilcoxon rank test. Data are presented as modified sina plots where boxes represent interquartile ranges and medians, and notches approximate 95% confidence intervals. f, Manhattan plots of mRNAs encoded on MMU16 differentially expressed by genotype in the mesenteric lymph nodes and colored as indicated. *q < 0.1 determined by DESeq2 with Benjamini-Hochberg correction. Source data

Extended Data Fig. 3. Triplication of the Ifnr locus drives increased IFNR protein expression and an aberrant antiviral response in a mouse model of Down syndrome.

a, Relative geometric mean fluorescent intensities (gMFIs) for IL10RB protein on CD45 + white blood cells (WBCs) from heterozygous Ifnr knockout mice (WT^1xIfnrs, n = 14, 3 male, 11 female), wild-type (WT, n = 17, 7 male, 10 female), Dp16 (n = 15, 7 male, 8 female), and Dp16^2xIfnrs (n = 15, 8 male, 7 female), by flow cytometry. b, Soluble IFNAR2 protein by ELISA in the plasma of each biological replicate relative to the average WT value per experiment. The number of animals per group are the following: n = 28 WT^1xIfnrs, n = 30 WT; n = 13 Dp16, and n = 26 Dp16^2xIfnrs (undocumented sex ratio). c, Example histogram for IFNR proteins on WBCs by flow cytometry. d, Pseudo-colored plot of gating strategy for immune subsets in whole blood by flow cytometry. Areas of high relative population density shown in red and orange, mid density in yellow, and low density in green and blue. e, IFNR protein expression on immune subsets isolated as in (d) from CD45 + WBCs. Heatmap indicates relative fluorescent intensities (RFIs) for gMFIs of the indicated genotype over the WT average IFNR gMFI per experiment. Gray denotes IFNRs not detected above isotype background. The number of animals is the same as for panel a. f, Representative histograms with gMFIs indicated for phosphorylated STAT1 (pSTAT1) by flow cytometry of WBCs at baseline or after stimulation with IFNα (10,000 U/mL) or IFNγ (100 U/mL) for 30 minutes. g, Cytokine levels in the plasma of mice treated with poly(I:C). Heatmap indicates log₂ fold-change (FC) of cytokine protein in plasma of the indicated cohort +poly(I:C) relative to the WT poly(I:C) cohort after 3 days of the poly(I:C) regimen. WT poly(I:C) (n = 7, 3 male, 4 female), Dp16 poly(I:C) (n = 6, 5 male, 1 female), and Dp16^2xIfnrs poly(I:C) (n = 6, 3 male, 3 female). In a-b, each dot represents an independent biological animal replicate with the mean indicated with a dash. For a-b, e, and g, significance (*p < 0.05) and exact p-values were determined by a two-sided Mann Whitney test. Source data

Extended Data Fig. 4. Correction of Ifnr locus copy number prevents embryonic heart malformations in a mouse model of Down syndrome.

a, Representative images of haematoxylin and eosin stained serial sections through entire murine hearts at embryonic day (E)15.5. Serial sections were cut through the entire region of the developing heart with transverse directionality indicated by arrow. Images represent a total of 58 formalin fixed paraffin embedded embryos that were processed and analyzed across four independent experiment batches; R: right, L: left, A: atrium, V: ventricle. WT (n = 24, 12 male, 12 female), Dp16 (n = 18, 9 male, 7 female, 2 undocumented sex), and Dp16^2xIfnrs (n = 16, 4 male, 12 female). b, Western blot analysis of total and phosphorylated STAT1 at tyrosine 701 (pSTAT1) protein from developing hearts at E15.5 of wild-type (WT), heterozygous Ifnr knockout mice (WT^1xIfnrs), Dp16, and Dp16^2xIfnrs animals, where n = 3/4/3/3 per group (undocumented sex ratio), respectively. c, Heatmaps displaying Normalized Enrichment Scores (NES) from Gene Set Enrichment Analysis (GSEA) of transcriptome changes in E18.5 and adult heart tissue, sorted by NES for Dp16 versus WT. E18.5 - WT (n = 6, 4 male, 2 female), Dp16 (n = 6, 3 male, 3 female), Dp16^2xIfnrs (n = 6, 3 male, 3 female), Adult - WT (n = 5, 2 male, 3 female), Dp16 (n = 6, 3 male, 3 female), Dp16^2xIfnrs (n = 5, 3 male, 2 female); asterisks indicate q < 0.1 defined by GSEA after Benjamini-Hochberg correction. d, Heatmaps displaying median RPKM expression Z-scores per genotype for example genes from gene sets dysregulated in E12.5 and adult heart tissue. e, Sina plots displaying expression levels in adult heart tissue for example genes. q-values determined by DESeq2 with significance set at q < 0.1 after Benjamini-Hochberg correction. In d-e, sample sizes are as described in c. Source data

Extended Data Fig. 5. Cognition and behavior are improved with corrected gene dosage of the Ifnr locus in a mouse model of Down syndrome.

a, Odds ratio plots for developmental milestone achievement in neonates as assessed by mixed effects Cox regression for the indicated pairwise comparisons between Dp16, Dp16^2xIfnrs, and wild-type (WT) animals, separated by sex (Female-F/Male-M), with adjustment for litter (random). Square points represent ‘success’ ratios with size proportional to -log₁₀(q), error bars indicate 95% confidence intervals, red indicates q < 0.1 after Benjamini-Hochberg correction, dashed lines indicate odds ratio=1. b, Sina plots showing distributions for day of developmental milestone achievement, with boxes indicating interquartile ranges and medians, and notches approximating 95% confidence intervals; sample sizes shown in panel a. c, Growth curves for neonates assessed in a-b. Data are represented as means ± SEM. p-values determined by repeated measures two-way ANOVA. d-e, Schematics of Morris water maze (MWM). Mice navigate to a hidden platform more efficiently over time (d). Swim sessions are divided into two phases where the platform is in opposite quadrants labeled by intercardinal directions (e). f, Duration (top) and total distance of path traveled (bottom) until platform escape during MWM for all animals or separated by sex; data are represented as means ± SEM, with significance determined by two-way repeated measures ANOVA with Tukey’s HSD test; asterisks indicate p < 0.05 and are colored by comparison: Dp16 versus WT (blue), Dp16^2xIfnrs versus WT (orange), Dp16 versus Dp16^2xIfnrs (red). See Source Data Extended Data Fig. 5 for all p-values. g, Distributions of average distance of mice from MWM center; horizontal dashes represent means. WT (n = 39, 19 male, 20 female), Dp16 (n = 28, 13 male, 15 female), and Dp16^2xIfnrs (n = 25, 12 male, 13 female). h, Time until fall from a rotating rod. WT (n = 16, 8 male, 8 female), Dp16 (n = 14, 7 male, 7 female), and Dp16^2xIfnrs (n = 17, 8 male, 9 female). For g-h, p-values were determined by one-way ANOVA with Tukey’s HSD test. i, Heatmap displaying Normalized Enrichment Scores (NES) from Gene Set Enrichment Analysis (GSEA) of transcriptome changes in adult brains, sorted by Dp16 versus WT (n = 6, 2 male, 4 female), Dp16 (n = 5, 2 male, 3 female), and Dp16^2xIfnrs (n = 7, 4 male, 3 female); asterisks indicate q < 0.1 by GSEA after Benjamini-Hochberg correction. Source data

Extended Data Fig. 6. Triplication of the Ifnr locus exacerbates craniofacial morphometric differences in a mouse model of Down syndrome.

a, Representative micro-computed tomography (μCT) reconstructions of a wild-type (WT) mouse skull. Landmarks are indicated on dorsal views of the outer skull proper (left) and interior cranial base (center) as well as on a lateral view of the hemi-mandible (right). b, Form difference ratio of inter-landmark distances for the skull proper and mandible by Euclidean Distance Matrix Analysis for cohorts of Dp16 and Dp16^2xIfnrs mice compared to WT controls, followed by bootstrapping 10,000x. The precise number of animals per group prior to bootstrapping are WT (n = 7, 2 male, 5 female), Dp16 (n = 7, 4 male, 3 female), and Dp16^2xIfnrs (n = 6, 4 male, 2 female). Data are represented as mean population estimates (black dots) with 95% confidence intervals (lines) colored according to differences versus WT where green represents larger inter-landmark distance, blue represents smaller inter-landmark distance, and gray represents confidence intervals that intersect 1 (that is, no difference). c, Distributions of form difference ratios (natural log (ln)-transformed) for all mean population estimates of inter-landmark distances on the skull proper and mandible shown in panel (b) that were smaller (blue) or larger (green) in WT versus Dp16 or versus Dp16^2xIfnrs (smaller, blue, or larger, green). WT (n = 7, 2 male, 5 female), Dp16 (n = 7, 4 male, 3 female), and Dp16^2xIfnrs (n = 6, 4 male, 2 female). p-values determined by two-sided Mann-Whitney test with significance set at p < 0.05. d, Scatter plots comparing mRNA fold-changes for Dp16 differentially expressed genes (DEGs) in facial mesenchyme tissue from E10.5 mice for Dp16/WT and Dp16^2xIfnrs/WT, with Ifnrs highlighted in red, Dp16 triplicated genes in blue, non-triplicated Dp16 DEGs in gray, and slope (m) colored accordingly; solid gray lines represent linear fits for the non-triplicated Dp16 DEGs. WT (n = 3, 2 male, 1 female), Dp16 (n = 3, 2 male, 1 female), Dp16^2xIfnrs (n = 3, 2 male, 1 female). e, Heatmap displaying Normalized Enrichment Scores (NES) from Gene Set Enrichment Analysis (GSEA) of transcriptome changes in E10.5 facial mesenchyme for indicated comparisons, sorted by the for Dp16 versus WT comparison; asterisks indicate q < 0.1 by GSEA after Benjamini-Hochberg correction.

Extended Data Fig. 7. Overview of experimental evidence regarding the genetic basis of heart defects in mouse models of Down syndrome.

Diagram depicts genetic variants on human chromosome 21 (HSA21) that may contribute to risk of congenital heart defects (CHDs) in humans with trisomy 21 (T21, yellow), genes with functional evidence whose triplication is necessary or sufficient to increase incidence of CHDs in mouse models of Down syndrome (DS, blue), and genes with supporting evidence in both humans and mouse models of DS (green)^,,,,–. Relative cytogenetic locations and number of bolded protein-coding genes are indicated along ideogram of the q arm of HSA21 colored according to Giemsa banding. Sources supporting this summary overview is provided in Source Data Extended Data Fig. 7.