Innate Immunity and Asthma Risk in Amish and Hutterite Farm Children The new england journal of medicine n engl j med 375;5 nejm.org August 4, 2016 411 established in 1812 August 4, 2016 vol. 375 no. 5 The authors’ affiliations are listed in the Appendix. Address reprint requests to Dr. Vercelli at the Arizona Respiratory Cen- ter, University of Arizona, The Bio5 Insti- tute, Rm. 339, 1657 E. Helen St., Tucson, AZ 85721, or at donata@ email . arizona . edu; to Dr. Ober at the Department of Human Genetics, University of Chicago, 920 E. 58th St., CLSC 425, Chicago, IL 60637, or at c-ober@ genetics . uchicago . edu; or to Dr. Sperling at the Department of Medi- cine, University of Chicago, 924 E. 57th St., JFK R316, Chicago, IL 60637, or at asperlin@ uchicago . edu. Ms. Stein, Dr. Hrusch, and Ms. Gozdz and Drs. von Mutius, Vercelli, Ober, and Sperling contributed equally to this article. This article was updated on August 4, 2017, at NEJM.org. N Engl J Med 2016;375:411-21. DOI: 10.1056/NEJMoa1508749 Copyright © 2016 Massachusetts Medical Society. BACKGROUND The Amish and Hutterites are U.S. agricultural populations whose lifestyles are remark- ably similar in many respects but whose farming practices, in particular, are distinct; the former follow traditional farming practices whereas the latter use industrialized farming practices. The populations also show striking disparities in the prevalence of asthma, and little is known about the immune responses underlying these disparities. METHODS We studied environmental exposures, genetic ancestry, and immune profiles among 60 Amish and Hutterite children, measuring levels of allergens and endotoxins and assess- ing the microbiome composition of indoor dust samples. Whole blood was collected to measure serum IgE levels, cytokine responses, and gene expression, and peripheral- blood leukocytes were phenotyped with flow cytometry. The effects of dust extracts obtained from Amish and Hutterite homes on immune and airway responses were assessed in a murine model of experimental allergic asthma. RESULTS Despite the similar genetic ancestries and lifestyles of Amish and Hutterite children, the prevalence of asthma and allergic sensitization was 4 and 6 times as low in the Amish, whereas median endotoxin levels in Amish house dust was 6.8 times as high. Differences in microbial composition were also observed in dust samples from Amish and Hutterite homes. Profound differences in the proportions, phenotypes, and func- tions of innate immune cells were also found between the two groups of children. In a mouse model of experimental allergic asthma, the intranasal instillation of dust ex- tracts from Amish but not Hutterite homes significantly inhibited airway hyperreactiv- ity and eosinophilia. These protective effects were abrogated in mice that were deficient in MyD88 and Trif, molecules that are critical in innate immune signaling. CONCLUSIONS The results of our studies in humans and mice indicate that the Amish environment provides protection against asthma by engaging and shaping the innate immune re- sponse. (Funded by the National Institutes of Health and others.) a b s t r a c t Innate Immunity and Asthma Risk in Amish and Hutterite Farm Children Michelle M. Stein, B.S., Cara L. Hrusch, Ph.D., Justyna Gozdz, B.A., Catherine Igartua, B.S., Vadim Pivniouk, Ph.D., Sean E. Murray, B.S., Julie G. Ledford, Ph.D., Mauricius Marques dos Santos, B.S., Rebecca L. Anderson, M.S., Nervana Metwali, Ph.D., Julia W. Neilson, Ph.D., Raina M. Maier, Ph.D., Jack A. Gilbert, Ph.D., Mark Holbreich, M.D., Peter S. Thorne, Ph.D., Fernando D. Martinez, M.D., Erika von Mutius, M.D., Donata Vercelli, M.D., Carole Ober, Ph.D., and Anne I. Sperling, Ph.D. The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016412 T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e Many genetic risk factors have been reported to modify susceptibility to asthma and allergy,1,2 but the dra- matic increase in the prevalence of these condi- tions in westernized countries in the past half- century suggests that the environment also plays a critical role.3 The importance of environmental exposures in the development of asthma is most exquisitely illustrated by epidemiologic studies conducted in Central Europe that show signifi- cant protection from asthma and allergic disease in children raised on traditional dairy farms. In particular, children’s contact with farm animals and the associated high microbial exposures4,5 have been related to the reduced risk.6,7 However, the effect of these traditional farming environ- ments on immune responses is not well defined. To address this gap in knowledge, we de- signed a study that compares two distinctive U.S. farming populations — the Amish of Indi- ana and the Hutterites of South Dakota — that recapitulate the differences in the prevalences of asthma and allergy observed in farmers and nonfarmers in Europe. These two particular groups of farmers originated in Europe — the Amish in Switzerland and the Hutterites in South Tyrol — during the Protestant Reformation and then emigrated to the United States in the 1700s and 1800s, respectively. Both groups have since remained reproductively isolated.8,9 Their lifestyles are similar with respect to most of the factors known to influence the risk of asthma, including large sibship size, high rates of child- hood vaccination, diets rich in fat, salt, and raw milk, low rates of childhood obesity, long dura- tions of breast-feeding, minimal exposure to tobacco smoke and air pollution, and taboos against indoor pets. However, whereas the Amish practice traditional farming, live on single- family dairy farms, and use horses for fieldwork and transportation, the Hutterites live on large, highly industrialized, communal farms. Strik- ingly, the prevalence of asthma in Amish versus Hutterite schoolchildren is 5.2% versus 21.3% and the prevalence of allergic sensitization is 7.2% versus 33.3%, as previously reported.10,11 M e t h o d s Overview We characterized the immune profiles of Amish and Hutterite schoolchildren. Furthermore, we used mouse models of asthma to study the effect of the environment on airway responses and to create a mechanistic framework for the interpre- tation of our observations in humans. Study Participants and Study Oversight In November 2012, we studied 30 Amish children 7 to 14 years of age who lived in Indiana, and in December 2012 we studied 30 Hutterite children who lived in South Dakota and were matched with the Amish children for sex and for age within 1 year. (For information on the characteristics of the children see Table 1, and the Methods section in the Supplementary Appendix, available with the full text of this article at NEJM.org.) Written informed consent was obtained from the parents and written assent was obtained from the chil- dren. One parent of each child responded to a questionnaire on asthma symptoms and previous diagnoses. The study was approved by the insti- tutional review boards at the University of Chi- cago and at St. Vincent Hospital in Indianapolis. Blood-Sample Collection and Analysis Whole blood was collected in tubes that con- tained culture medium alone, medium plus 0.1 μg per milliliter of lipopolysaccharide, or medium plus 0.4 μg per milliliter of anti-CD3 plus 0.33 μg per milliliter of anti-CD28 monoclonal antibod- ies (TruCulture Blood Collection System, Myriad RBM). After incubation at 37°C for 30 hours, A Quick Take is available at NEJM.org Characteristic Amish (N = 30) Hutterite (N = 30) Age (yr) Median 11 12 Range 8–14 7–14 Girls (no.) 10 10 Sibships (no.) 15 14 Children with asthma (no.) 0 6 Positivity for allergen-specific IgE (no.) >0.7 kUA/liter 5 9 >3.5 kUA/liter 2 9 Serum IgE (kU/liter) Median 21 64 Interquartile range 10–57 15–288 * UA denotes allergen-specific unit. Table 1. Demographic and Clinical Characteristics of the Study Populations.* The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016 413 Immunit y a nd A s thm a in A mish a nd Hu t ter ite Childr en supernatant and cells were frozen for use in gene-expression and cytokine studies. Levels of 26 cytokines were measured with the use of the Milliplex Map Human TH17 Magnetic Bead Panel (EMD Millipore) or enzyme-linked immuno- absorbent assay (eBioscience) in the supernatant, in accordance with standard protocols. Addi- tional blood was collected to obtain peripheral- blood leukocytes for flow cytometry and DNA isolation, and serum was collected for IgE stud- ies (as described in the Methods section in the Supplementary Appendix). Cryopreserved human peripheral-blood leuko- cytes were incubated for 10 minutes with pooled human IgG antibodies (FcX, Biolegend) to block nonspecific antibody binding before undergoing surface staining with f luorescently conjugated antibodies (see Table S1 in the Supplementary Appendix) and intracellular staining for FoxP3 (eBioscience). Flow-cytometry data were acquired on an LSRFortessa cell analyzer (BD Biosciences), and acquisition data were analyzed with FlowJo software (Tree Star). Dust Collection and Extract Preparation Electrostatic dust collectors were placed in one bedroom and the living room in each of 10 Amish and 10 Hutterite homes to collect air- borne house dust. All 10 Amish homes and 9 of 10 Hutterite homes housed children who partici- pated in the study. After 1 month, dust was ana- lyzed for endotoxin and allergen levels, and ex- tracts were prepared for studies in mice. In addition, a vacuum was used to collect dust from the living-room floor in Amish homes and from mattresses in Amish and Hutterite homes for use in microbiome studies (as described in the Methods section in the Supplementary Appendix). Aqueous extracts of house dust from Amish and Hutterite homes were prepared as described in the Methods section in the Supplementary Appendix. Genetic Studies RNA was extracted from thawed cells with the use of AllPrep DNA/RNA Mini Kits (Qiagen). RNA underwent complementary DNA synthesis and was then hybridized to HumanHT-12 v4 Expression BeadChip arrays (Illumina). A com- mon set of 118,789 single-nucleotide polymor- phisms (SNPs) was genotyped or imputed in the 60 children in the study (see Fig. S1 in the Sup- plementary Appendix). Mouse Models We instilled 50 μl of house-dust extract intra- nasally every 2 to 3 days (for a total of 14 times) into 7-week old BALB/c mice (Harlan Laborato- ries), beginning on day 0. The mice had been sensitized intraperitoneally with 20 μg of oval- bumin (grade V, Sigma) plus alum (Pierce) on days 0 and 14 and were challenged intranasally with 50 μg of ovalbumin on days 28 and 38. Beginning 5 days before day 0, we also instilled 50 μl of dust extract from Amish homes intra- nasally every 2 to 3 days (for a total of 14 times) in 7-week old, C57BL6 wild-type, MyD88-deficient mice12 and in mice deficient in both MyD88 and Trif13 (Jackson Laboratories). These mice were sensitized intraperitoneally with 20 μg of oval- bumin plus alum on days 0 and 14 and chal- lenged intranasally with 75 μg of ovalbumin on days 26, 27, and 28. Statistical Analysis Statistical analyses were performed with the use of R or Prism software (GraphPad). Differences in the distributions of median cytokine levels between Amish and Hutterite children were de- termined with use of a Wilcoxon signed-rank test. We used linear regression to identify differ- entially expressed genes in the Amish and Hut- terite untreated samples of peripheral-blood leukocytes. The methods of Benjamini and Hoch- berg14 were used to control the false discovery rate. For flow cytometric studies and the study in mice, differences in cell populations and air- way resistance were assessed with the use of an unpaired Student’s t-test. Additional details on sample processing, quality control, and statisti- cal analysis for all methods described here are provided in the Methods section in the Supple- mentary Appendix. R e s u l t s Asthma and Allergic-Sensitization Rates and Genetic Ancestry None of the Amish children and six (20%) of the Hutterite children had asthma, rates similar to those reported in earlier studies.10,11 Levels of total serum IgE and the number of children whose levels of IgE against common allergens were high (defined as more than 3.5 kUA [allergen- specific unit] per liter) were lower in the Amish group than in the Hutterite group (Table 1, and The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016414 T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e Table S2 in the Supplementary Appendix). No statistical differences were observed in levels of serum Ig isotypes other than IgE (Fig. S2 in the Supplementary Appendix). To evaluate whether the differences in the prevalence of asthma and of allergic sensitiza- tion and differences in IgE level could be attrib- uted to population history, we assessed ancestry by conducting principal component analysis and compared allele frequencies using genomewide SNPs. These studies revealed remarkable genetic similarities between the Amish children and the Hutterite children, as compared with other Euro- pean populations15 (Fig. 1A, and Fig. S3 in the Supplementary Appendix). Exposures to Allergens, Microbes, and Microbial Products Common allergens (from cats, dogs, house-dust mites, and cockroaches) were detectable in air- borne dust from 4 of 10 Amish and 1 of 10 Hut- terite homes (Table S3 in the Supplementary Appendix). In contrast, endotoxin levels were measurable in airborne dust from all 20 homes, and median levels were strikingly higher (6.8 times as high) in Amish homes than in Hutterite homes (4399 endotoxin units [EU] per square meter vs. 648 EU per square meter, P <0.001) (Fig. 1B). Analysis of a single pooled sample of mattress dust from each population revealed different profiles of the relative abundance of bacteria at the family level (Fig. S4 in the Supplementary Appendix). Composition and Phenotype of Peripheral- Blood Leukocytes Peripheral-blood leukocytes from Amish children had increased proportions of neutrophils, de- creased proportions of eosinophils, and similar proportions of monocytes as compared with samples from Hutterite children (Fig. 2A). Neu- trophils from Amish children expressed lower levels of the chemokine receptor CXCR4 and the adhesion molecules CD11b and CD11c than did neutrophils from Hutterite children, suggesting that these cells may have recently emigrated from the bone marrow (Fig. 2B). Although propor- tions of monocytes were similar in Amish and Hutterite children, monocytes from Amish chil- dren, unlike those from Hutterite children, ex- hibited a suppressive phenotype characterized by lower levels of human leukocyte antigen DR (HLA-DR) and higher levels of the inhibitory molecule immunoglobulin-like transcript 3 (ILT3)16,17 (Fig. 2C). In contrast with previous studies,18,19 no significant differences in percent- Figure 1. Ancestries and Environments of Amish and Hutterite Children. Panel A shows a principal components plot of the first two principal com- ponents (PC 1 and PC 2) of the analysis of 72,034 single-nucleotide poly- morphisms (SNPs). Amish and Hutterite genotypes were projected onto the sample space created by Human Genome Diversity Project (HGDP) for European populations.15 Panel B shows endotoxin levels in airborne dust from 10 Amish and 10 Hutterite homes. Box-and-whisker plots show a hor- izontal line indicating median value, a box representing the interquartile range, and whiskers showing the 95% confidence interval. The P value was calculated with the use of the Wilcoxon rank-sum test. EU denotes endo- toxin units. A SNP Analysis of Genetic Association P C 2 20 0 −20 −40 −30 −20 −10 0 10 20 30 PC 1 B Endotoxin Levels in Airborne Dust E U /m 2 Amish Homes Hutterite Homes 20,000 10,000 15,000 5,000 0 P<0.001 Amish Hutterite Basque French North Italian Russian Russian Caucasus Sardinian Scottish Tuscan The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016 415 Immunit y a nd A s thm a in A mish a nd Hu t ter ite Childr en Figure 2. Proportions of Peripheral-Blood Leukocytes and Cell-Surface–Marker Phenotypes in Amish and Hutterite Children. The percentages of total peripheral-blood leukocytes (Panel A) were determined with flow cytometry for neutrophils (defined as CD66b+CD16+), eosinophils (defined as CCR3+Siglec-8+), and monocytes (defined as CD14+CD66b−). Box-and-whisker plots show a line indicating median value, with the box showing the interquartile range and whiskers showing the 95% confidence interval. Neutrophils (Panel B) were characterized according to the surface expression of CXCR4, CD11b, and CD11c (shown here), along with CXCR1 and CXCR2, expressed as mean fluorescence intensity (MFI). The expression of the interleukin-8 coreceptors CXCR1 and CXCR2 was not signifi- cantly different between groups (P = 0.26 and P = 0.91, respectively). Monocytes (Panel C) were characterized for the surface expression of HLA-DR and immunoglobulin-like transcripts (ILTs), including ILT3 (shown here). There was no significant difference in the MFI of in- hibitory receptors ILT2 and ILT4 between Amish and Hutterite children (P = 0.69 and P = 0.21, respectively; data not shown), whereas the surface expression of ILT5 was increased on Amish monocytes (P = 0.001; data not shown). All P values were calculated with the use of an unpaired Student’s t-test. Cell proportions and phenotypes after the exclusion of children with asthma or allergic sensitization are shown in Table S4 in the Supplementary Appendix. C D 66 b + C D 16 + ( % ) 80 60 40 0 Amish Hutterite B Cell-Surface Markers on Neutrophils A Cell Proportions of Peripheral-Blood Leukocytes Neutrophils P=0.006 C C R 3+ S ig le c- 8+ ( % ) 12 6 9 3 0 Amish Hutterite Eosinophils P<0.001 C D 14 + C D 66 b – ( % ) 2 3 1 0 Amish Hutterite Monocytes P=0.28 M F I 600 400 200 0 Amish Hutterite CXCR4 P<0.001 M F I 15,000 9,000 12,000 6,000 3,000 0 Amish Hutterite CD11b P<0.001 M F I 600 400 300 100 500 200 0 Amish Hutterite CD11c P=0.004 C Cell-Surface Markers on Monocytes M F I 4000 3000 1000 2000 0 Amish Hutterite HLA-DR P=0.004 M F I 2000 1500 2500 500 1000 0 Amish Hutterite ILT3 P=0.004 The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016416 T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e ages of T regulatory cells (defined as CD3+, CD4+, FoxP3+, and CD127−) was observed in Amish and Hutterite children (0.056±0.054% vs. 0.079±0.081% of peripheral-blood leukocytes, P = 0.29). Cytokine Responses to Innate and Adaptive Stimulation Cytokine levels were measured in supernatants from peripheral-blood leukocytes that were cul- tured for 30 hours, with or without innate stimuli (lipopolysaccharide) or adaptive stimuli (combined anti-CD3 and anti-CD28 antibodies). Twenty-three cytokines were detectable in the supernatants from peripheral-blood leukocytes treated with lipopolysaccharide. Median levels of each of these 23 cytokines were lower in the Amish children than in the Hutterite children, and these distributions were significantly differ- ent (P <0.001 by Wilcoxon signed-rank test) (see Tables S5 and S6 in the Supplementary Appen- dix). Results were similar after the exclusion of children who were known to have asthma or allergies (Table S7 in the Supplementary Appen- dix). In contrast, after adaptive stimulation, the overall distributions of median cytokine levels were not significantly different in peripheral- blood leukocytes from Amish and Hutterite children (P = 0.08 by Wilcoxon signed-rank test) (Table S8 in the Supplementary Appendix). Gene-Expression Profiles The striking differences in the proportions of peripheral-blood leukocytes observed in Amish and Hutterite children were reflected in the gene- expression profiles of these cells (Fig. S5 in the Supplementary Appendix). At a false discovery rate of 1%, 1449 genes were up-regulated in the peripheral-blood leukocytes of Amish children (blue points in Fig. 3A) as compared with 1360 genes up-regulated in the cells of Hutterite chil- dren (red points in Fig. 3A). These differentially expressed genes were organized into 15 coexpres- sion modules with the use of the Whole Genome Co-Expression Network Analysis (Table S9 in the Supplementary Appendix). To better understand the biologic relationships within each set of genes in each module, we used Ingenuity Pathway Analysis (Qiagen) to construct unsupervised net- works on the basis of prior knowledge of the physical and functional connections between the molecules encoded by the genes. The most sig- nificant network (P = 1.0×10−30 by Fisher’s exact test) was in a module that contained 43 genes. This module was associated with both Amish and Hutterite status (P = 7.1×10−9) and was therefore also associated with the proportions of neutro- phils (P = 1.5×10−6) and eosinophils (P = 1.0×10−3). Eighteen of the genes in this module were over- expressed in Amish peripheral-blood leukocytes, and all were clustered in a network that had as hubs tumor necrosis factor (TNF) and interferon regulatory factor 7 (IRF7), two key proteins in the innate immune response to microbial stimuli (Fig. 3B). Effects of House-Dust Extracts on Experimental Asthma To create a framework that would help us to interpret our observations, which were pointing toward a protective role of innate immunity, we used a classic ovalbumin mouse model of aller- gic asthma, comparing the effects of house dust obtained from Amish and Hutterite homes by administering extracts intranasally to mice over the course of 4 to 5 weeks. Eosinophilia was observed in bronchoalveolar-lavage samples, and airway hyperresponsiveness was exacerbated in mice treated with ovalbumin and Hutterite dust extracts as compared with mice treated with ovalbumin alone, findings that were consistent with the absence of protection from asthma observed in Hutterite children (Fig. 4A). In con- trast, inhalation of Amish dust extracts was suf- ficient to significantly inhibit ovalbumin-induced airway hyperresponsiveness, eosinophilia in the bronchoalveolar lavage, and the elevation of se- rum ovalbumin-specific IgE levels (Fig. 4A, and Table S10 in the Supplementary Appendix). Levels of lung T regulatory cells (defined as CD3+, CD4+, and FoxP3+) were not increased (Table S11 in the Supplementary Appendix), and all cyto- kines measured in bronchoalveolar-lavage sam- ples, including interleukin-10, were suppressed in mice that received Amish dust extracts (Table S12 in the Supplementary Appendix). The inhibi- tory effects of these extracts in wild-type mice probably required innate immunity, because protection was strongly reduced in mice defi- cient in MyD88 (Fig. 4C) and completely abro- gated in mice deficient in both MyD88 and Trif (Fig. 4D), two molecules that are critical to the development of multiple innate immune-signal- ing pathways. The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016 417 Immunit y a nd A s thm a in A mish a nd Hu t ter ite Childr en D i s c u s s i o n Our studies in Amish and Hutterite schoolchil- dren revealed marked differences in the preva- lence of asthma despite similar genetic ances- tries and lifestyles. As compared with the Hutterites, the Amish, who practice traditional farming and are exposed to an environment rich in microbes, showed exceedingly low rates of asthma and distinct immune profiles that sug- gest profound effects on innate immunity. Data generated in an experimental model of asthma support this notion by showing that the protec- tive effect of the Amish environment requires the activation of innate immune signaling. Analyses of the proportions and gene-expres- sion profiles of peripheral-blood immune cells in Amish and Hutterite children revealed differences in the cells and genes involved in innate immune responses to microbes. Indeed, neutrophils, eosino- phils, and monocytes appeared to be major tar- gets of the distinct environments to which Amish and Hutterite children are exposed be- cause these cell types differed between the two groups in terms of their relative abundance, their phenotypes, or both. Moreover, the network most associated with these differences consisted of innate immune genes. Notable among the genes that were more highly expressed in the Amish children was TNFAIP3, which encodes A20, a ubiquitin-editing enzyme that limits the activity of multiple inflammatory pathways that depend on nuclear factor κB (NF-κB)20 and that has also been shown to mediate the protective effects of European farm-dust extracts in murine models of allergic asthma.21 IRF7, a hub in this network, Figure 3. Gene-Expression Profiles in Peripheral-Blood Leukocytes from Amish and Hutterite Children. In Panel A, a volcano plot shows differences in baseline gene expression in peripheral-blood leukocytes from Amish and Hutterite children.The x axis indicates the log2 differences in gene-expression level between groups, with larger positive values representing genes with higher expression in the Hutterites relative to the Amish (1360 genes, shown in red points) and larger negative values representing genes with higher expression in the Amish relative to the Hut- terites (1449 genes, shown in blue points). The y axis shows the –log10 of the P values for each gene, with larger values indicating greater statistical significance. The solid horizontal line indicates the 1% false discovery rate. Black points represent genes from Amish and Hutterite cells for which there was no significant difference in gene expression. Differences in gene expression remain after the data for children with asthma or allergic sensitization were excluded (Figs. S6 and S7 in the Supplementary Appendix). Changes in gene expression between the two groups after correct- ing for differences in cell proportion are shown in Figure S4 in the Supplementary Appendix. In Panel B, a network of differentially expressed genes in untreated peripheral-blood leukocytes is shown. Genes shown in blue have in- creased expression in Amish children, and the gene shown in red has increased expression in Hutterite children. The gene shapes indicate the class of each gene’s protein product (spirals denote enzymes, a v-shape denotes cyto- kines, conjoined circles denote a transcription regulator, hollow upside-down triangles denote kinases, cups denote transporters, and circles denote other products). Lines represent different biologic relationships (solid lines indicate direct interaction, dashed lines indirect interaction, arrows direction of activation, arrows with a horizontal line direc- tion of activation and inhibition, and lines without arrows binding only). 13601449 STEAP4 ZC3H12A IRAK3 TRIM25 TRIM8 TNFAIP3 PARP14 TAP2 PARP12 SAMD9L PSME1 IRF7 DHX58 MAP3K8 STAT2TNFAIP2 SCO2 RHBDF2 TNF 15 10 5 0 −2 −1 0 Log2 (difference in gene expression) − Lo g 1 0 (P v al u e) 1 2 BA The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016418 T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e F ig u re 4 . E ff ec ts o f A m is h a n d H u tt er it e H o u se -D u st E xt ra ct s o n A ir w ay R es p o n se s in M o u se M o d el s o f A lle rg ic A st h m a. P an el A s h o w s th e ef fe ct s o f th e in tr an as al i n st ill at io n o f 50 μ l o f A m is h o r H u tt er it e d u st e xt ra ct i n 7 -w ee k- o ld m ic e (B A LB /c s tr ai n ) ev er y 2 to 3 d ay s fo r a to ta l o f 14 t im es b eg in - n in g at d ay 0 . T h e m ic e w er e se n si ti ze d w it h o va lb u m in ( O V A ) in tr ap er it o n ea lly o n d ay s 0 an d 1 4 an d c h al le n g ed w it h o va lb u m in i n tr an as al ly o n d ay s 2 8 an d 3 8 . A ir w ay r es is ta n ce (s h o w n a s ce n ti m et er s o f w at er p er m ill ili te r p er s ec o n d a n d s ti m u la te d i n r es p o n se t o i n cr ea si n g d o se s o f ac et yl ch o lin e ad m in is te re d i n tr av en o u sl y) a n d b ro n ch o al ve o la r- la va g e (B A L) c el lu la ri ty w er e m ea su re d o n d ay 3 9 (4 t o 6 m ic e p er g ro u p ). T h e to ta l am o u n t o f A m is h a n d H u tt er it e d u st e xt ra ct a d m in is te re d o ve r th e co u rs e o f th e ex p er im en t re p re se n t- ed t h e to ta l lo ad o f ai rb o rn e d u st d ep o si te d o n e le ct ro st at ic d u st c o lle ct o rs p la ce d i n A m is h o r H u tt er it e h o m es f o r 1 m o n th . S ta ti st ic al d if fe re n ce s in e xp er im en ta l m ea su re s w er e as se ss ed w it h t h e u se o f S tu d en t’ s t- te st . A m is h h o u se -d u st e xt ra ct s (7 .5 m g o f d u st e q u iv al en t in 5 0 μ l) w er e in st ill ed i n tr an as al ly e ve ry 2 t o 3 d ay s fo r a to ta l o f 14 t im es b eg in n in g 5 d ay s b ef o re d ay 0 i n to 7 -w ee k o ld w ild -t yp e m ic e (P an el B ), m ic e d ef ic ie n t in M yD 8 8 (P an el C ), a n d m ic e d ef ic ie n t in M yD 8 8 an d T ri f (P an el D ) (a ll C 57 B L6 s tr ai n s) . T h es e m ic e w er e se n si ti ze d i n tr ap er it o n ea lly w it h 2 0 μ g o f o va lb u m in o n d ay s 0 an d 1 4 an d w er e ch al le n g ed i n tr an as al ly w it h 7 5 μ g o f o va lb u m in o n d ay s 26 , 27 , an d 2 8 . A ir w ay r es is ta n ce (s h o w n a s an i n cr ea se f ro m b as el in e in r es p o n se t o i n cr ea si n g d o se s o f n eb u liz ed m et h ac h o lin e) a n d b ro n ch o al ve o la r- la va g e ce llu la ri ty w er e m ea su re d o n d ay 3 0 (1 2 m ic e p er g ro u p f o r w ild -t yp e m ic e an d 6 m ic e p er g ro u p f o r th o se d ef ic ie n t in M yD 8 8 o r M yD 8 8 an d T ri f) . S ta ti st ic al d if fe re n ce s in e xp er im en ta l m ea su re s w er e as se ss ed w it h t h e u se o f S tu - d en t’ s t- te st . I b ar s re p re se n t th e st an d ar d e rr o rs o f th e d at a. N S d en o te s n o t si g n if ic an t an d P B S p h o sp h at e- b u ff er ed s al in e. S al in e O V A O V A –A m is h S al in e O V A O V A –A m is h S al in e O V A O V A –A m is h P B S O V A O V A – A m is h O V A – H u tt er it e Airway Resistance (cm of water/ml/sec) 20 1015 5 0 0 1. 00 0. 50 0. 25 2. 00 4. 00 A ce ty lc h o lin e (µ g/ g m o u se ) B W ild T yp e C 57 B L6 C C 57 B L6 M yD 88 K n o ck O u t D C 57 B L6 M yD 88 –T R if K n o ck O u t A B A LB /c P = 0. 04 P = 0. 02 P = 0. 01P < 0. 00 1 P < 0. 00 1 P = 0. 00 7 Airway Resistance (cm of water/ml/sec) 10 68 4 2 0 0 10 30 10 0 M et h ac h o lin e (m g/ m l)P = 0. 03 Airway Resistance (cm of water/ml/sec) 10 68 4 2 0 0 10 30 10 0 M et h ac h o lin e (m g/ m l)N S Airway Resistance (cm of water/ml/sec) 10 68 4 2 0 0 10 30 10 0 M et h ac h o lin e (m g/ m l)N S BAL Cells (%) 10 0 5075 25 BAL Cells (%) 10 0 5075 25 BAL Cells (%) 10 0 5075 25 BAL Cells (%) 10 0 5075 25 0 E os in op hi ls N eu tro ph ils M ac ro ph ag es P < 0. 00 1 0 0 P = 0. 02 P = 0. 01 P < 0. 00 1 P = 0. 00 8 P < 0. 00 1 N S N S 0 N S P = 0. 00 2 N S Eo sin op hi ls N eu tro ph ils M ac ro ph ag es Eo sin op hi ls N eu tro ph ils M ac ro ph ag es Eo sin op hi ls N eu tro ph ils M ac ro ph ag es P B S O V A O V A – A m is h O V A – H u tt er it e S al in e O V A O V A –A m is h S al in e O V A O V A –A m is h S al in e O V A O V A –A m is h The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016 419 Immunit y a nd A s thm a in A mish a nd Hu t ter ite Childr en regulates type I interferon transcription and is therefore essential for innate airway responses against viruses22 that are linked to susceptibility to asthma.23,24 In turn, TRIM8, the one gene in the network that was more highly expressed in the Hutterites, acts as a positive regulator of TNF-α– and interleukin-1β–induced activation of NF-κB.25 These findings suggest that in the Amish, intense and presumably sustained exposure to microbes activates innate pathways that shape and cali- brate downstream immune responses. Sustained microbial exposure was also reflect- ed in the phenotypes of peripheral innate im- mune cells in the Amish. Repeated microbial stimulation can lead to reduced expression of HLA-DR on monocytes26,27 and drive immature neutrophils from the bone marrow.28-31 Indeed, Amish children had immature neutrophils bear- ing markers suggestive of recent emigration from the bone marrow, and they had monocytes with reduced expression of HLA-DR and in- creased expression of ILT3, all of which are sug- gestive of antiinflammatory function. Proportions of T regulatory cells and levels of interleukin-10, which typically mediate immune-balancing effects, were not increased in the Amish children. How- ever, qualitative and functional differences in regulatory-cell populations remain to be defined. Innate immunity has evolved to sense the environment and transduce signals that cali- brate adaptive responses to exogenous antigens. The proteins MyD88 and Trif are located at the convergence of multiple innate signaling path- ways,32 and deletion of these molecules virtually disables innate immune responses, thereby also dysregulating adaptive immunity. The fact that the loss of protection was more marked in mice deficient in both MyD88 and Trif than in mice deficient only in MyD88 points to the involve- ment of multiple innate pathways. The concor- dance between findings from studies in humans and in mice was remarkable: in both studies protection was accompanied by lower levels of eosinophils, higher levels of neutrophils, general- ly suppressed cytokine responses, and no increase in levels of T regulatory cells or interleukin-10. Thus, the finding that these features were largely dependent on innate immune pathways in mice suggests that innate immune signaling may also be the primary target of protection in the Amish children, in whom downstream adaptive immune responses may also be modulated. Our study has several limitations. First, we were unable to include children younger than 6 years of age, we collected samples at a single time point, and the numbers of Amish and Hutterite children in our study were relatively small. As a result, we may have missed important windows of immune development or lacked the ability to detect early, subtle shifts in cell composition, response, or phenotype that are critical for im- mune maturation. Second, our microbiome as- sessments were limited, since only pooled dust samples from a limited number of homes were available for the studies in which we assessed bacterial composition. Therefore, we cannot fur- ther dissect microbial composition and identify potentially protective microbes to target. How- ever, the striking differences found in endotoxin levels support the notion that the Amish indoor environment is much richer in microbial expo- sures than the Hutterite environment. Third, the strategy used for sampling the Hutterite chil- dren enriched selection for those with asthma, although the prevalence of asthma in our sample was similar to that reported in previous popula- tion-based studies.11 Moreover, the exclusion of children with asthma or allergic sensitization from our analyses of gene expression, cell com- position, and immune phenotypes did not affect the outcomes. Our study in a small number of children was sufficient to show significant differences in the prevalence of asthma and in immune profiles, suggesting that very strong environmental fac- tors must account for these differences. Indeed, we showed that there are remarkable genetic similarities between Amish and Hutterite chil- dren. Although we interrogated only common variants, other variants that occur at very low frequency in these populations are unlikely to account for the observed large differences in the prevalence of asthma. In the end, the novelty of our work lies in the identification of innate im- munity as the primary target of the protective Amish environment, a finding supported by re- sults obtained in both humans and mice. Con- versely, our work suggests that susceptibility to asthma may be increased when innate immune stimulation is weak. A deeper understanding of the relevant stimuli and the innate immune path- ways they engage may ultimately pave the way for the development of effective strategies for the prevention of asthma. The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016420 T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e Supported by the National Institutes of Health, St. Vincent Foundation, and the American Academy of Allergy, Asthma, and Immunology Foundation. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank the Hutterite and Amish volunteers and their families for participating in this study; Gorka Alkorta-Aranburu, Maitane Arrubarrena Orbegozo, Kathleen Bailey, Christine Billstrand, Kelly Blaine, Daniel Cook, Donna Decker, Mohammad Jaffery, Courtney Burrows, Katherine Naughton, Raluca Nicolae, Rob Stanaker, Meghan Sullivan, and Emma Thompson for assistance on field trips and with sample processing; Peace Ezeh, Amanda Herrell, Ashley Horner, Kenneth Addison, Dominik Schenten, and Shane Snyder for their contributions to the studies in mice; and Minal Çalışkan, Yoav Gilad, Jessie Nicodemus-Johnson, John Novembre, and Matthew Stephens for helpful comments and statistical advice. Appendix The authors’ affiliations are as follows: the Department of Human Genetics (M.M. Stein, C.I., R.L.A., C.O.), the Department of Medi- cine, Section of Pulmonary and Critical Care Medicine, and the Committee on Immunology (C.L.H., A.I.S.), the Department of Ecology and Evolution (J.A.G.), and the Department of Surgery (J.A.G.), University of Chicago, Chicago, and the Institute for Genomic and Systems Biology, Argonne National Laboratory, Argonne (J.A.G.) — all in Illinois; the NIEHS Training Program in Environmental Toxicology and Graduate Program in Cellular and Molecular Medicine (J.G.), and the Departments of Cellular and Molecular Medicine (V.P., D.V.), Medicine (J.G.L.), Chemical and Environmental Engineering (M. Marques dos Santos), and Soil, Water, and Environmental Science (J.W.N., R.M.M.), University of Arizona, and the Arizona Respiratory Center and Bio5 Institute (J.G., V.P., S.E.M., J.G.L., F.D.M., D.V.) — all in Tucson; the Department of Occupational and Environmental Health, University of Iowa, Iowa City (N.M., P.S.T.); Allergy and Asthma Consultants, Indianapolis (M.H.); and Dr. von Hauner Children’s Hospital, Ludwig Maximilians University Munich, Munich, Germany (E.M.). References 1. Ober C, Yao TC. The genetics of asth- ma and allergic disease: a 21st century perspective. Immunol Rev 2011; 242: 10-30. 2. Meyers DA, Bleecker ER, Holloway JW, Holgate ST. Asthma genetics and per- sonalised medicine. Lancet Respir Med 2014; 2: 405-15. 3. Bach JF. The effect of infections on susceptibility to autoimmune and allergic diseases. N Engl J Med 2002; 347: 911-20. 4. Braun-Fahrländer C, Riedler J, Herz U, et al. Environmental exposure to endo- toxin and its relation to asthma in school- age children. N Engl J Med 2002; 347: 869- 77. 5. Ege MJ, Mayer M, Normand AC, et al. Exposure to environmental microorgan- isms and childhood asthma. N Engl J Med 2011; 364: 701-9. 6. Ege MJ, Frei R, Bieli C, et al. Not all farming environments protect against the development of asthma and wheeze in children. J Allergy Clin Immunol 2007; 119: 1140-7. 7. Ege MJ, Strachan DP, Cookson WO, et al. Gene-environment interaction for childhood asthma and exposure to farm- ing in Central Europe. J Allergy Clin Im- munol 2011; 127: 138-44 8. Hostetler JA. Hutterite society. Balti- more: Johns Hopkins University Press, 1974. 9. Hostetler JA. Amish society. 4th ed. Baltimore: Johns Hopkins University Press, 1993. 10. Holbreich M, Genuneit J, Weber J, Braun-Fahrländer C, Waser M, von Mutius E. Amish children living in northern Indi- ana have a very low prevalence of allergic sensitization. J Allergy Clin Immunol 2012; 129: 1671-3. 11. Motika CA, Papachristou C, Abney M, Lester LA, Ober C. Rising prevalence of asthma is sex-specific in a US farming population. J Allergy Clin Immunol 2011; 128: 774-9. 12. Adachi O, Kawai T, Takeda K, et al. Targeted disruption of the MyD88 gene results in loss of IL-1- and IL-18-mediated function. Immunity 1998; 9: 143-50. 13. Yamamoto M, Sato S, Hemmi H, et al. Role of adaptor TRIF in the MyD88-inde- pendent toll-like receptor signaling path- way. Science 2003; 301: 640-3. 14. Benjamini Y, Hochberg Y. Control- ling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 1995; 57: 289-300. 15. Cann HM, de Toma C, Cazes L, et al. A human genome diversity cell line panel. Science 2002; 296: 261-2. 16. Kim-Schulze S, Scotto L, Vlad G, et al. Recombinant Ig-like transcript 3-Fc modu- lates T cell responses via induction of Th anergy and differentiation of CD8+ T sup- pressor cells. J Immunol 2006; 176: 2790-8. 17. Velten FW, Duperrier K, Bohlender J, Metharom P, Goerdt S. A gene signature of inhibitory MHC receptors identifies a BDCA3(+) subset of IL-10-induced den- dritic cells with reduced allostimulatory capacity in vitro. Eur J Immunol 2004; 34: 2800-11. 18. Lluis A, Depner M, Gaugler B, et al. Increased regulatory T-cell numbers are associated with farm milk exposure and lower atopic sensitization and asthma in childhood. J Allergy Clin Immunol 2014; 133: 551-9. 19. Schaub B, Liu J, Höppler S, et al. Ma- ternal farm exposure modulates neonatal immune mechanisms through regulatory T cells. J Allergy Clin Immunol 2009; 123(4): 774-82.e5. 20. Ma A, Malynn BA. A20: linking a complex regulator of ubiquitylation to immunity and human disease. Nat Rev Immunol 2012; 12: 774-85. 21. Schuijs MJ, Willart MA, Vergote K, et al. Farm dust and endotoxin protect against allergy through A20 induction in lung epithelial cells. Science 2015; 349: 1106-10. 22. Bosco A, Ehteshami S, Panyala S, Martinez FD. Interferon regulatory factor 7 is a major hub connecting interferon- mediated responses in virus-induced asthma exacerbations in vivo. J Allergy Clin Immunol 2012; 129: 88-94. 23. Loss GJ, Depner M, Hose AJ, et al. The early development of wheeze: environ- mental determinants and genetic suscep- tibility at 17q21. Am J Respir Crit Care Med 2016; 193: 889-97. 24. Calışkan M, Bochkov YA, Kreiner- Møller E, et al. Rhinovirus wheezing ill- ness and genetic risk of childhood-onset asthma. N Engl J Med 2013; 368: 1398-407. 25. Li Q, Yan J, Mao AP, et al. Tripartite motif 8 (TRIM8) modulates TNFα- and IL-1β-triggered NF-κB activation by tar- geting TAK1 for K63-linked polyubiquiti- nation. Proc Natl Acad Sci U S A 2011; 108: 19341-6. 26. Meisel C, Schefold JC, Pschowski R, et al. Granulocyte-macrophage colony- stimulating factor to reverse sepsis-asso- ciated immunosuppression: a double-blind, randomized, placebo-controlled multi- center trial. Am J Respir Crit Care Med 2009; 180: 640-8. 27. Gomez HG, Gonzalez SM, Londoño JM, et al. Immunological characterization of compensatory anti-inf lammatory re- sponse syndrome in patients with severe sepsis: a longitudinal study. Crit Care Med 2014; 42: 771-80. 28. Fortunati E, Kazemier KM, Grutters JC, Koenderman L, Van den Bosch J. Hu- man neutrophils switch to an activated phenotype after homing to the lung irre- spective of inflammatory disease. Clin Exp Immunol 2009; 155: 559-66. The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved. n engl j med 375;5 nejm.org August 4, 2016 421 Immunit y a nd A s thm a in A mish a nd Hu t ter ite Childr en 29. Orden S, De Pablo C, Rios-Navarro C, et al. Efavirenz induces interactions be- tween leucocytes and endothelium through the activation of Mac-1 and gp150,95. J Antimicrob Chemother 2014; 69: 995-1004. 30. Devi S, Wang Y, Chew WK, et al. Neu- trophil mobilization via plerixafor-medi- ated CXCR4 inhibition arises from lung demargination and blockade of neutro- phil homing to the bone marrow. J Exp Med 2013; 210: 2321-36. 31. Yamada M, Kubo H, Kobayashi S, et al. The increase in surface CXCR4 expres- sion on lung extravascular neutrophils and its effects on neutrophils during endotoxin- induced lung injury. Cell Mol Immunol 2011; 8: 305-14. 32. Netea MG, Wijmenga C, O’Neill LAJ. Genetic variation in toll-like receptors and disease susceptibility. Nat Immunol 2012; 13: 535-42. Copyright © 2016 Massachusetts Medical Society. Tucson, Arizona Gregory Koshkarian, M.D. The New England Journal of Medicine Downloaded from nejm.org at CARNEGIE-MELLON UNIV on April 5, 2021. For personal use only. No other uses without permission. Copyright © 2016 Massachusetts Medical Society. All rights reserved.