A specific primed immune response in Drosophila is dependent on phagocytes

Abstract

Drosophila melanogaster, like other invertebrates, relies solely on its innate immune response to fight invading microbes; by definition, innate immunity lacks adaptive characteristics. However, we show here that priming Drosophila with a sublethal dose of Streptococcus pneumoniae protects against an otherwise-lethal second challenge of S. pneumoniae. This protective effect exhibits coarse specificity for S. pneumoniae and persists for the life of the fly. Although not all microbial challenges induced this specific primed response, we find that a similar specific protection can be elicited by Beauveria bassiana, a natural fly pathogen. To characterize this primed response, we focused on S. pneumoniae-induced protection. The mechanism underlying this protective effect requires phagocytes and the Toll pathway. However, activation of the Toll pathway is not sufficient for priming-induced protection. This work contradicts the paradigm that insect immune responses cannot adapt and will promote the search for similar responses overlooked in organisms with an adaptive immune response.

Figure 1. Protection from a Priming Dose of S. pneumoniae Persists for the Life of the Fly

(A) Survival curves of flies injected with PBS (triangles, n = 60) or 250 (circles, n = 60) or 3,500 (squares, n = 65) CFU S. pneumoniae. p < 0.001, comparing 3,500 CFU to the other treatments (log-rank analysis). (B) Flies were injected with 250 (circles) or 3,500 (squares) CFU of S. pneumoniae. Bars represent geometric means of bacterial load with 95% confidence intervals. (C) Flies were primed on day 0 with PBS (circles), 250 CFU of S. pneumoniae (triangles), or dead S. pneumoniae (squares). One week later, flies were injected again with either PBS (open shapes) or 3,000 CFU S. pneumoniae (filled shapes). Naïve PBS-injected flies challenged with 3,000 CFU of S. pneumoniae die significantly faster than flies primed with 250 CFU of S. pneumoniae or dead S. pneumoniae (p < 0.0001, log-rank test). Dotted lines correspond to double injection controls. n = 158 to 228 for each condition. (D) Flies were injected with PBS (circles) or dead S. pneumoniae (squares) on day 0 and challenged 1 wk later with 400 CFU of S. pneumoniae. Bars represent geometric means of bacterial load with 95% confidence intervals. (E) Flies were primed on day 0 with PBS (open bars) or dead S. pneumoniae (filled bars) and challenged with 5,000 CFU of S. pneumoniae on the indicated days. Mean survival with standard error values are plotted. n = 37 to 49 for each condition. At each time point, the survival curves differ significantly. See Figure S1 for log-rank analysis and individual survival curves.

Figure 2. The S. pneumoniae–Induced Primed Response Is Specific for S. pneumoniae

(A) Flies were injected with a priming dose of PBS (white bars), the same bacteria used for the lethal challenge (gray bars), or S. pneumoniae (black bars). Bacteria used for lethal challenges are indicated below the graph. Mean survival with standard error values are plotted. n = 57 to 128 for each condition. p < 0.001 for the indicated set of bars (log-rank test). See Figure S2 for individual survival curves and log-rank analysis. (B) Flies were primed on day 0 with PBS (circles, n = 59), dead S. pneumoniae (squares, n = 60), mixture 1 (dead E. coli, M. luteus, and B. bassiana; triangles, n = 59), or mixture 2 (dead E. coli, M. luteus, B. bassiana, and S. pneumoniae; triangles, n = 60) and challenged 1 wk later with 3,000 CFU of S. pneumoniae. Log-rank analysis indicates that curves corresponding to PBS and mixture 1 (dotted lines) are significantly different from those for flies primed with S. pneumoniae or mixture 2 (solid lines) (p < 0.001).

Figure 3. The Toll Pathway Is Required for the Primed Response

Partial loss-of-function alleles of PGRP-SA (circles) and imd (triangles) were injected with PBS (open shapes) or dead S. pneumoniae (filled shapes) on day 0 and challenged 1 wk later with a lethal dose of S. pneumoniae (20 CFU for PGRP-SA^seml and 100 CFU for imd¹⁰¹⁹¹). Log-rank analysis of the survival curves indicates that naïve and S. pneumoniae–primed PGRP-SA^seml flies die at the same rate, whereas the curves for naïve and S. pneumoniae–primed imd¹⁰¹⁹¹ flies are significantly different (p < 0.0001). n = 154 to 245 for each condition. Molecular information for imd¹⁰¹⁹¹ is given in Materials and Methods.

Figure 4. S. pneumoniae–Primed Flies Exhibit an Enhanced Phagocytic Response That Is Specific to S. pneumoniae

(A) Model for S. pneumoniae–induced primed response. (B and C) RNA was extracted from whole flies. Defensin transcript levels were quantified using qRT-PCR and normalized to 0-h media injection. Bars represent mean values with standard deviation. See Figures S4 and S5 for diptericin and attacin transcript levels. (B) Flies were injected with media (open bars), 250 CFU of S. pneumoniae (gray bars), or 250 CFU of E. coli (black bars). (C) Flies were primed on day 0 with media (white bars) or dead S. pneumoniae (gray bars) and challenged 1 wk later with media, 3,500 CFU of S. pneumoniae, or 3,500 CFU of E. coli (indicated above the graph). (D) Flies were injected with polystyrene beads (triangles, n = 122) or water (circles, n = 115) 3 d prior to day 0 to fully inhibit phagocytosis. On day 0, flies were injected with 20 CFU of S. pneumoniae. Bead-inhibited flies die significantly faster than do mock-injected flies (p < 0.0001, log-rank analysis of survival curves). (E) Flies were injected with polystyrene beads (circles) or water (squares) on day 0, injected with a priming dose of PBS (open shapes) or S. pneumoniae (filled shapes) on day 3, and injected with a lethal dose of 1,000 CFU on day 10. Log-rank analysis of the survival curves indicates that all flies died significantly faster than did mock-treated S. pneumoniae–primed flies (p < 0.0001). n = 51 to 62 for each condition. (F) Flies were injected with PBS (circles) or dead S. pneumoniae (squares) on day 0 and challenged 1 wk later with 6,000 CFU of E. coli. Bars represent geometric means of bacterial load with 95% confidence intervals.