Pathogen evasion of social immunity

Abstract

Treating sick group members is a hallmark of collective disease defence in vertebrates and invertebrates alike. Despite substantial effects on pathogen fitness and epidemiology, it is still largely unknown how pathogens react to the selection pressure imposed by care intervention. Using social insects and pathogenic fungi, we here performed a serial passage experiment in the presence or absence of colony members, which provide social immunity by grooming off infectious spores from exposed individuals. We found specific effects on pathogen diversity, virulence and transmission. Under selection of social immunity, pathogens invested into higher spore production, but spores were less virulent. Notably, they also elicited a lower grooming response in colony members, compared with spores from the individual host selection lines. Chemical spore analysis suggested that the spores from social selection lines escaped the caregivers' detection by containing lower levels of ergosterol, a key fungal membrane component. Experimental application of chemically pure ergosterol indeed induced sanitary grooming, supporting its role as a microbe-associated cue triggering host social immunity against fungal pathogens. By reducing this detection cue, pathogens were able to evade the otherwise very effective collective disease defences of their social hosts.

Fig. 1. Pathogen selection under individual versus social immunity.

a, Serial passage experiment of fungal pathogens (green) in ants in the absence (left) or presence (right) of caregiving nestmates. Spores collected from carcasses (ants with spore outgrowth) were used for the next infection cycle. b, Proportion of the three M. robertsii (R1–3) and three M. brunneum (B1–3) strains still present from the starting mix after five (top) and ten (bottom) host passages for the individual (I1–10) and social (S1–10) replicate lines. The diversity of strains present per line decreased slower under social immunity (passage 5: WRST, P = 0.015; Supplementary Table 1), but reached equally low values of a single remaining or dominating strain per line under both selection treatments at the end of the experiment (passage 10: WRST, P = 1.000, Supplementary Table 1). M. brunneum was more successful under social immunity. At the end of the experiment, M. robertsii strain R1 was dominant in 7 individual and 6 social lines, strain R3 in 3 individual and 2 social lines, and M. brunneum strain B2 in 2 social lines, so that the final composition of strains after selection did not differ between the two treatments (Fisher’s exact test, P = 0.629; Supplementary Table 1). All statistics based on 10 individual and 10 social lines.

Fig. 2. Effect of social immunity on pathogen virulence and transmission.

a, The presence of caregivers in the common garden experiment reduced pathogen virulence, measured as induced host mortality, for the individual lines, while the social lines showed overall low virulence, which was not further reduced by nestmate presence (Bayesian MLM, contrasts Ind-Ind to all other groups < 0.02, see Supplementary Table 1). b, Social lines produced about double the amount of spores per carcass than the individual lines, independent of current nestmate presence (LMM, P < 0.001, Supplementary Table 1). Dots depict the 20 replicate lines, each tested in two current host conditions (R1-dominant lines indicated by enlarged symbols; for separate statistics see Supplementary Table 2); black lines indicate means and shaded areas indicate 95% CI. Letters denote posthoc differences of P < 0.05.

Fig. 3. Pathogen evasion of social immunity.

a, Social lines elicited less allogrooming in nestmates (WRST, P = 0.003; 20 replicate lines, each tested in 3 biological replicates of 30 min videos each; colours as in Fig. 2). b, Their chemical profiles show reduction in six compounds (blue with stars), mostly representing ergosterol (insert; WRST, P = 0.025, Supplementary Table 1; dots represent each of the 10 replicate lines, each measured in 3 technical replicates; shaded area: 95% CI around medians (black line)) and its derivatives (Supplementary Fig. 2 and Extended Data Fig. 1; all other fungal compounds black with arrows, compound numbers as in Supplementary Table 3; non-fungal compounds grey, internal standards as filled peaks in dark grey). Abundance in chromatogram given in counts per scan ×10⁷, in ergosterol insert as ISTD response factor. c,d, Application of pure ergosterol (blue) induced increased nestmate allogrooming over the sham control (c) (WRST, P = 0.026, Supplementary Table 1; n = 22 ergosterol- and 23 sham-treated replicates, each in 30 min video), whereas this was not the case for the structurally similar, yet not fungus-derived cholesterol (d) (grey; WRST, P = 0.144, Supplementary Table 1; n = 23 cholesterol- and 24 sham-treated replicates, each in 30 min video). Violin- and boxplots in a, c and d show replicate lines as dots, medians as lines within the box defined by the interquartile range, and whiskers revealing minimum to maximum range; kernel density presented as shade in violin plots. Enlarged dots in a and b indicate R1-dominant lines. Letters denote posthoc differences of P < 0.05 between selection histories; **P < 0.01 vs the sham control; NS, non-significant. Details for all statistical analyses are provided in Supplementary Table 1 (for the subset of R1-dominant lines only, see Supplementary Table 2).

Extended Data Fig. 1. Abundance of the five additional spore compounds affected by social immunity.

Further to ergosterol (compound 18; Fig. 3b insert), five additional compounds were identified by the random forest as important in differentiating between the two selection histories (shown in the order of their derived compound importance; Supplementary Fig. 2). Three of these (compounds 16, 17, 19) were classified as ergosterol derivatives, one as an ester (compound 25) and one is unknown (compound 7; for mass spectrum see Supplementary Fig. 3). All of the identified important compounds were significantly lower in abundance in the social than the individual lines (WRST, comp. 7 and 17, p = 0.007; comp. 19, p = 0.025; comp. 25 and 16: p = 0.031; Supplementary Table 1; significant differences denoted by different letters). Dots represent the 10 replicate lines each of the individual and social selection history (with enlarged dots indicating R1-dominant lines; for separate statistics see Supplementary Table 2), shades the 95% CIs around the medians (black lines). Colours as in Fig. 2.