Lunar-linked biological rhythms in the immune system of freshwater three-spined stickleback

Abstract

Immune responses are widely accepted to be under circadian regulation via a molecular clock, with many practical consequences, but much less is known of how other biological rhythms could affect the immune system. In this study, we search for lunar rhythms (circalunar, circasemilunar, and circatidal cycles) in the immune expression of the recently marine-derived freshwater fish, the low-plate morph of the three-spined stickleback. We employed time series of immune expression (mRNA) measurements for 14 immune-associated genes, representing a variety of immunological pathways. Times series measurements were taken on fish populations in the wild, in seminatural outdoor mesocosms, and in the laboratory, according to sampling regimens originally designed to study circannual variation but with the additional potential to provide information about lunar variation. Our evidence best supported the existence of a very small endogenous tidal rhythm. This is consistent with previous suggestions of the existence of a primordial tidal endogenous clock, some elements of which may be conserved in animals evolving outside the marine environment.

Keywords: immunity; lunar; three-spined stickleback; tidal; ultradian.

Graphical Abstract

Figure 1:

scatterplots of the physical quantities representing the different cycles considered in this study against time (epoch 00:00 h 1 January 1900); points represent individual sampling occasions in the field sampling regimen (river, lake, and mesocosms). Annual variation is represented by photoperiod; lunar variation is represented by the proportion of the lunar disc illuminated; semilunar variation is represented by tidal amplitude on the adjacent coastline; and tidal variation is represented by tidal height on the adjacent coastline

Figure 2:

matrix pie chart plots for the field sampling regimen (river, lake, and mesocosms) showing Pearson correlations between the cosinor terms used to represent different cycles in overall analyses and also the physical quantity for each cycle (annual variation = photoperiod; lunar variation = proportion of the lunar disc illuminated; semilunar variation = tidal amplitude on the adjacent coastline; tidal variation = tidal height on the adjacent coastline). Clockwise-orientated segments represent positive correlations

Figure 3:

scatterplots of the physical quantities representing the different cycles considered in this study against time (epoch 00:00 h 1 January 1900); points represent individual sampling occasions in the laboratory sampling regimen. Annual variation is represented by photoperiod; lunar variation is represented by the proportion of the lunar disc illuminated; semilunar variation is represented by tidal amplitude on the adjacent coastline; tidal variation is represented by tidal height on the adjacent coastline

Figure 4:

matrix pie chart plots for the laboratory sampling regimen, showing Pearson correlations between the cosinor terms used to represent different cycles in overall analyses and also the physical quantity for each cycle (annual variation = photoperiod; lunar variation = proportion of the lunar disc illuminated; semilunar variation = tidal amplitude on the adjacent coastline; tidal variation = tidal height on the adjacent coastline). Clockwise-orientated segments represent positive correlations

Figure 5:

heat maps indicate the percentage of variation in the expression of individual genes explained by different cycles (left panel) and the overall significance (across all genes) of the cycles (right panel) for the river, lake, mesocosm, and laboratory populations. Var, % of total variance explained in each gene in a LMM; P, overall P value for the cosinor terms in an MLMM. In the case of the lunar cycles for the field sites, the % total variance explained is scaled by the residual variation from the circannual base model

Figure 6:

timing of individually significant gene expression cycles where there was overall significance in a given habitat. Clockwise circular plots showing the acrophase of gene expression maxima in radians zeroed on the acrophase for the corresponding physical cycle (the highest amplitude spring tide for semicircalunar cycles and high tide for tidal cycles). (A) semicircalunar cycles at field sites. (B) Circatidal cycles in the laboratory

Figure 7:

smoothed density scatterplot of gene relative expression (RE) partial residuals against time (decimal tidal phase point) with line showing the predicted tidal sinusoid from an LMM. (A) RE for il1r1-like. (B) RE for gpx4a