Bed bugs live in an ecologically and spatially restricted indoor habitat comprised of overlapping aggregation and host odors, and they traverse relatively short distances between blood-hosts and aggregation sites. Although many studies demonstrated aggregation or host odor preference respectively, the modulation of their preferences between these divergent odors is poorly understood. Given the recurrent transitions of bed bugs between replete and hungry states, we evaluated the effects of six odorscapes containing aggregation and host skin odors on bed bug preferences. Hunger state modulated odor preference for aggregation and foraging in all tested odorscapes. Aggregation odor attracted both fed and unfed bed bugs. Host skin odor attracted unfed bed bugs but repelled recently fed bed bugs and the addition of carbon dioxide to host odor enhanced the behavioral responses. These findings suggest that orientation to aggregation sites in fed bed bugs is driven by two distinct odor processing mechanisms for attractant and aversive odors. Unfed bed bugs discriminate two attractive odors – aggregation and host odors – but host odor predominates over aggregation odor in driving their orientation behavior. Understanding the dynamic switching of odor preferences during the blood digestion cycle will guide the implementation of chemical lures in IPM. Host odors alone and their co-emission with aggregation pheromone repelled fed bed bugs from traps. Conversely, unfed bed bugs had a strong preference for host odor emitted either alone or with aggregation odors. Therefore, the independent use of either host or aggregation odor lures and their co-emission from the same trap should be carefully considered.

2          MATERIALS AND METHODS

2.1       Insects

The tested laboratory strain (Harold Harlan strain) was collected in 1973 in Fort Dix NJ, and maintained on a human host until December 2008. Then, in our laboratory, it was fed on defibrinated rabbit blood until July 2021 and on human blood thereafter. It was maintained at 35–45% relative humidity, 25 °C, on a 12:12 (L:D) h cycle and fed weekly on heparinized human blood (supplied by the American Red Cross under IRB #00000288 and protocol #2018-026). We used an artificial feeding system, which has been previously described. The feeding system was housed in a North Carolina State University-approved BSL-2 facility (Biological Use Authorization #2020-09-836). Between feeding sessions, the glass feeders were sanitized with 7.5% sodium hypochlorite and 95% ethanol, and air-dried. Since we conducted 24 hr-observations in an open arena, only adult males (age unknown) were used in this study in case of accidental escape. Within 24 hrs post feeding, fully engorged males were separated from colony jars into groups of 100. Each group was kept in a new rearing jar (5 cm diameter x 4.5 cm height) with two clean filter paper shelters (4 x 9 cm) using #1 Whatman filter papers (Whatman, Madistone, United Kingdom) for 2-, 4-, 6-, 8-, 10-, 12- or 14-days. Tested males were discarded after a single bioassay.

2.2       Open arena bioassay design

Arena bioassays evaluated the preferences of bed bugs for human skin odor and aggregation odor. The bottom of a plastic arena (50L x 40W x 15H cm, without a lid, #2565519, Project Source, Baton Rouge, LA) was lined with disposable absorbent liner (AL2050, Jaece Industries, North Tonawanda, NY), secured with masking tape (2020-48TP6, 3M Scotch, Maplewood, MN), and replaced after each trial. A pitfall shelter was placed on each side of the arena, 35 cm apart (Fig. 1). Pitfall shelters were made as follows: Commercial disposable bathroom paper cups (88.7 mL, 5 cm diameter x 5.6 cm height) were placed below the arena and attached to the disposable absorbent liner so there was no gap between the cup and the liner (Fig. 1). A filter paper with a test odor (see below) was placed in each cup and the cup was lightly covered with clean filter papers during the assay to create a dark shelter within the cup. Under the dark conditions in the scotophase, the shelter did not appear to serve as a visual landmark, but it could potentially be visually detected by bed bugs during the photophase. Importantly, the shelters did not serve as traps and therefore we refer to them as pitfall shelters rather than pitfall traps; the vertical test filter paper touched the horizontal cover filter paper, so the tested bed bugs could move freely between the two shelters during each bioassay. However, we saw no evidence of bed bugs leaving a shelter after they entered it.

2.3       Preparation of odor sources

Bed bug-conditioned filter papers, human skin swabs and clean filter papers were used as the aggregation odor, host skin odor and no-odor sources, respectively. We did not carry out chemical analyses because we tested natural blends of skin and aggregation odors. For preparing the aggregation odor, two filter papers, each 4 x 9 cm made from #1 Whatman filter papers, were placed in a jar with 100 freshly fed males for 14 days. The filter papers served as substrate and shelter on which the bed bugs defecated. Because we used three arenas for each treatment, the two conditioned filter papers were cut into pieces, and an equal amount of filter paper was placed in each pitfall shelter. Thus, a single pitfall shelter contained 24 cm² of bed bug-conditioned filter paper, representing approximately 33-male-equivalents of aggregation odor.

For preparing the host odor, human skin was swabbed with two filter papers (#1 Whatman) following a previously validated protocol with IRB approval for recruitment, informed consent, and odor sample collections granted from North Carolina State University, Raleigh, NC (IRB Approval #14173). A single adult male participant (>21 yr old) provided informed consent. Prior to collecting skin swabs, he was instructed to: 1) not eat ‘spicy’ food at least 24 h before collecting a skin swab; 2) take a morning shower; 3) not use a deodorant or cosmetics/lotions on the sampled surfaces; 4) not exercise or perform any strenuous physical activity; and 5) take the skin swabs 4–8 h after showering. The participant was then provided with #1 Whatman filter papers and glass vials (20 ml) and asked to collect skin swabs as follows: 1) rinse hands with water and dry before use; 2) use a single filter paper and swab the left arm from hand to armpit for 12 s using both sides of the filter paper; 3) rub the left leg from the lower thigh to ankle for 12 s using both sides of the filter paper; 4) rub the left armpit for 6 s using both sides of the filter paper; 5) place the filter paper into a glass vial and label the vial; and 6) repeat with a new filter paper swabbing the right side of the body. The papers were kept at –30 ^oC and tested within 24 hrs after collection. Three skin swab papers were cut into pieces, which were equally distributed among three pitfall shelters in three replicates. Thus, a single pitfall shelter contained the equivalent of a single filter paper swabbed on a human subject for 30 sec. In Experiments 4 and 5 (see below), one pitfall shelter contained both aggregation and host odors represented by separate filter papers for aggregation odor and host odor.

2.4       Bioassay observations

The bioassay room was kept at 25 °C, 30–40% RH, and a photoperiod of 12 hrs dark (08:00 to 20:00) and 12 hrs light (20:00 to 08:00). These environmental conditions were the same as in the rearing environment. Bioassays were conducted for 24 hrs starting at 09:00. Twenty-five males (ages unknown) were placed under an inverted cup in the center of the arena (release point, Fig. 1) and allowed to acclimate for 3 min. The cup was removed, and the positions of all bed bugs (the two pitfall shelters and arena floor) were recorded at 0, 1, 3, 6, 12 and 24 hrs (09:00, 10:00, 12:00, 15:00, 21:00 and 09:00). Throughout, the human observer wore a face mask and gloves to minimize bed bug exposure to human breath and skin odors, including CO₂. Red light was used for observation during the scotophase. To avoid odor contamination among different treatments in the bioassay room, only one type of treatment (experiment) was conducted using three bioassay arenas on the same day.

2.5       Experimental design

We generated six types of odorscapes for six experiments, with three arenas used as three replicates in each experiment (Table 1).

Experiment 1. Both pitfall shelters had the same odor sources (either aggregation odor or host (skin) odor) to evaluate the symmetry of the bioassay environment. Two-days post feeding (Fed) and 14-days post feeding (Unfed) adult males were bioassayed.

Experiment 2. A narrow odorscape in which bed bugs were given a choice between a single odor source and no odor: either aggregation or host odor vs. clean filter paper. An odor source was placed in one pitfall shelter and clean filter paper in the other shelter. Two-days post feeding (Fed) and 14-days post feeding (Unfed) adult males were bioassayed.

Experiment 3. A broader odorscape in which bed bugs were given a choice between two distinct odor sources: aggregation odor vs. host odor. To evaluate the effect of hunger state on odor preference, separate cohorts of bed bugs were assayed 2-, 4-, 6-, 8-, 10-, 12- and 14-days post feeding.

Experiment 4. A complex odorscape in which bed bugs could choose between two admixed odor sources and a single odor source: a combination of aggregation and host odors in one pitfall shelter vs. host odor in the other shelter. Two-days post feeding (Fed) and 14-days post feeding (Unfed) adult males were bioassayed.

Experiment 5. In an inverse experiment to Experiment 4, a complex odorscape was tested in which bed bugs were given a choice between two admixed odor sources and a single odor source: a combination of aggregation and host odors in one pitfall shelter vs. aggregation odor in the other shelter. Two-days post feeding (Fed) and 14-days post feeding (Unfed) adult males were bioassayed.

Experiment 6. To add complexity to the odorscape, Experiment 5 was repeated, but with CO₂ added to the combination of aggregation odor and host odor in one pitfall shelter vs. aggregation odor in the other shelter. Because CO₂ is an effective host cue in bed bugs, we added CO₂ to host odor to mimic more complex and realistic host cues. As shown in Fig. 1, in this experiment only, both pitfall shelters received clean air (Medical quality air, Airgas Healthcare, Radnor, PA) that was passed through a humidifying jar (100 mL/min for each shelter). Carbon dioxide (2% or 20,000 ppm, Airgas Healthcare, Radnor, PA) was added to the pitfall shelter that contained the aggregation and host skin odors. To avoid the accumulation of CO₂ in the arena, observation time was limited to 0, 1, 3 and 6 hrs, all in the scotophase.

2.6       Statistical analysis

Binary choice odor preferences across multiple treatment groups were tested by full factorial repeated measures ANOVA (α = 0.05) with Tukey’s HSD in JMP (Student edition 18, Cary, NC), which enabled multiple comparisons of main (fixed) effects including the number of bed bugs in each of the two pitfall shelters (Pitfall shelter main effect) across all time points (Time main effect) and the interaction of these two factors (Pitfall shelter * Time). Notably, in all except one experiment, there was a significant effect of Time (P ≤ 0.05). Therefore, in the main manuscript we report the P-values of only the main effect (Pitfall shelter). The complete test results (F-values and df) for both main effects (Pitfall shelter, Time) and their interaction (Pitfall shelter * Time) are reported in the Supplementary Material. When a significant interaction effect was detected, we implemented Tukey’s HSD post-hoc test, and significant differences between the two pitfall shelters by time (P ≤ 0.05) are indicated by an asterisk (*) on the figures and reported in the Supplementary Material.

The effects of different odorscapes on odor preferences were compared using the numbers of bed bugs recorded in pitfall shelters at the 6, 12 and 24 hr time points and analyzed them with one-way ANOVA and Tukey’s HSD.