Multiple stages of evolutionary change in anthrax toxin receptor expression in humans

Abstract

The advent of animal husbandry and hunting increased human exposure to zoonotic pathogens. To understand how a zoonotic disease may have influenced human evolution, we study changes in human expression of anthrax toxin receptor 2 (ANTXR2), which encodes a cell surface protein necessary for Bacillus anthracis virulence toxins to cause anthrax disease. In immune cells, ANTXR2 is 8-fold down-regulated in all available human samples compared to non-human primates, indicating regulatory changes early in the evolution of modern humans. We also observe multiple genetic signatures consistent with recent positive selection driving a European-specific decrease in ANTXR2 expression in multiple tissues affected by anthrax toxins. Our observations fit a model in which humans adapted to anthrax disease following early ecological changes associated with hunting and scavenging, as well as a second period of adaptation after the rise of modern agriculture.

Fig. 1. Decreased ANTXR2 in humans can affect anthrax toxin sensitivity in blood cells.

a B. anthracis toxins lethal factor (LF) and edema factor (EF) cause apoptosis, edema, and immune suppression in target host cells. b Differentially transcribed protein-coding genes between humans (H) and non-human primates (NHP) in CD4 + T cells based on PRO-seq data (differential expression based on DESeq2, P < 0.0001, multiple testing correction made). The GO term “integral component of the membrane” is enriched in differentially transcribed genes, as seen in purple. ANTXR2 is highlighted in red. c) CD4 + T-cell PRO-seq data shows that ANTXR2 is transcribed eightfold lower in human than chimpanzee and rhesus macaque (P < 0.0001, DESeq2, multiple testing correction made). d Comparison of ANTXR2 RNA-seq expression levels in CD4 + T cells among a large set of humans (n = 91) to rhesus macaque and baboon. Human variation in expression does not overlap with rhesus macaque or baboon expression. Human individuals are primarily self-reported as white or Asian (see “Methods” for details). e CRISPRa induction in K562 cells results in significantly increased ANTXR2 expression after 24 and 48 h (P = 0.007, one-sided t test) measured using qRT-PCR. Data are presented as mean values + /− SEM. n = 3 independent CRISPRa inductions. Source data are provided as a Source Data file. f K562 cells that overexpress endogenous ANTXR2 are significantly less viable after an anthrax toxin challenge as measured by the Alamar blue viability stain (P = 0.002, one-sided t test, SE error bars). Data are presented as mean values + /− SEM. n = 6 independent CRISPRa inductions over two independent experiments. Source data are provided as a Source Data file.

Fig. 2. Changes in ANTXR2cis-regulatory element activity and chromatin structure.

a Genome browser shot of CD4 + T-cell PRO-seq (normalized by reads per million, plus strand in red, minus strand in blue) in human, chimpanzee, and rhesus macaque. Regulatory elements predicted by dREG are shown in teal. Cis-regulatory elements (CREs) tested in this study (CRE1-9) are shown below the data tracks. CREs having decreased activity in human (as determined by luciferase in (d)) are shaded. Overlap with validated CRISPR targets that result in reduced ANTXR2 expression are shown below the CREs in purple. b Hi-C/Micro-C in CD4 + T cells from human and rhesus macaque. Normalized contacts are shown in the heatmap scale to the left. ANTXR2 is located in the same topological domain (TAD) as the upstream gene PRDM8 in both species. TADs are marked with white dotted lines. Focal contacts with the ANTXR2 promoter are circled with black dotted lines. Zoomed-out gene-level Hi-C contacts are pictured below, with the region shown in the top panel outlined in white. c Virtual 4C-seq signal calculated based on the number of Hi-C contacts for CREs tested in the luciferase assay. CREs with decreased activity in human are colored. The anchor symbol at each CRE represents the bait region that all contacts are derived from. The arrow at the end of the arc denotes the ANTXR2 promoter. d Luciferase assay performed in Jurkat cells to test the activity of regulatory elements in human, chimpanzee, and rhesus macaque shows lower activity in human compared with either chimpanzee or rhesus macaque for CRE1, CRE2, CRE3, CRE4, and CRE9. Data are presented as mean values + /− SEM. One-sided t tested used and significant P values are shown in the figure. n = 3 independent transfections in each species. Source data are provided as a Source Data file.

Fig. 3. Selective sweep affects ANTXR2 expression in multiple tissues.

a Sweepfinder2 CLR (composite-likelihood-ratio) values for CEU (Northern and Western Europe), CHB (China), JPT (Japan), and YRI (Yoruba) populations show a predicted selective sweep in the CEU population upstream of ANTXR2 at 81–81.05 Mb. b Pi diversity values for CEU, CHB, JPT, and YRI populations show decreased diversity in the predicted selective sweep in CEU, CHB, and JPT (see Supplementary Fig. 17 for percentiles). c ENCODE DNase-I-seq data for T cells, muscle, heart, stomach, and lung show different numbers and locations of DNase-I hypersensitive sites, which indicates differential regulatory landscapes in cell types within the predicted selective sweep in CEU. d Differential expression of GTEx data for all available tissues showed a range of fold changes of individuals of European descent compared to individuals of African descent (as determined by mitochondrial haplotype). The majority of tissues showed greater expression in individuals of African descent (n = 22/27). Source data are provided as a Source Data file.

Fig. 4. Genetic differentiation in European populations consistent with multiple phases of selection driving changes in ANTXR2.

a Weighted Fst (fixation index) comparisons between CEU (Northern and Western Europe) and CHB (China), JPT (Japan), and YRI (Yoruba) HapMap populations at the ANTXR2 locus show a peak around ANTXR2. b PRO-seq (normalized by reads per million) and dREG signal (pictured in teal) at cis-regulatory element 2 (CRE2). rs41407844 falls within a dREG peak shared by chimpanzee and rhesus macaque and the ancestral allele G is conserved within the primate lineage. Conservation of CRE2 between humans and non-human primate tracks show human-specific single nucleotide polymorphisms (SNPs), human/chimpanzee differences (SNPs in black, insertion/deletions (INDELS) in red), human/non-human primate differences (SNPs in black, INDELS in red), and 100-way conservation. Net synteny tracks show the position of regions that have one-to-one orthologs in the chimpanzee and rhesus macaque genomes. c Martchenko et al. showed that CEU lymphoblastoid cell lines (LCLs) have lower sensitivity to anthrax toxins than LCLs derived from other populations. Data are presented as mean values + /− SD (CEU n = 60, CHB n = 43, JPT n = 44, YRI n = 84). The allele frequency of SNP rs41407844 across populations. Source data are provided as a Source Data file. d Genome-wide global Fst distribution. The SNP rs41407844 is above the 99.57 percentile of Fst of all SNPs genome-wide. e Rs41407844 is above the 98% percentile for pairwise Fst comparisons between CEU and CHB, JPT, and YRI HapMap populations genome-wide. f Model demonstrating the pattern of human migration (dark green) compared to the spread of Neolithic culture (light green) and anthrax disease distribution (brown).