€¦ · web viewintergenerational effects of nutrition on immunity: a systematic review and...
TRANSCRIPT
Intergenerational effects of nutrition on immunity: a systematic
review and meta-analysis
Catherine E. Grueber1,2,*, Lindsey J. Gray1,3, Katrina M. Morris4, Stephen J.
Simpson1,3 and Alistair M. Senior3,5
1School of Life and Environmental Sciences, Faculty of Science, University of Sydney, NSW
2006, Australia
2San Diego Zoo Global, PO Box 120551, San Diego, CA 92112, USA
3Charles Perkins Centre, University of Sydney, NSW 2006, Australia
4Roslin Institute, The University of Edinburgh, Easter Bush Campus, Midlothian EH25 9RG,
UK
5School of Mathematics and Statistics, Faculty of Science, University of Sydney, NSW 2006,
Australia
Running header: Intergenerational effects of nutrition on immunity
*Author for correspondence (E-mail: [email protected]; Tel.: +61 2 8627
0278).
ABSTRACT
Diet and immunity are both highly complex processes through which organisms interact with
their environment and adapt to variable conditions. Parents that are able to transmit
information to their offspring about the prevailing environmental conditions have a selective
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
advantage by ‘priming’ the physiology of their offspring. We used a meta-analytic approach
to test the effect of parental diet on offspring immune responses. Using the geometric
framework for nutrition (a method for analysing diet compositions wherein food nutrient
components are expressed as axes in a Cartesian coordinate space) to define dietary
manipulations in terms of their energy and macronutrient compositions, we compiled the
results of 226 experiments from 38 published papers on the intergenerational effects of diet
on immunity, across a range of study species and immunological responses. We observed
intergenerational impacts of parental nutrition on a number of offspring immunological
processes, including expression of pro-inflammatory biomarkers as well as decreases in anti-
inflammatory markers in response to certain parental diets. For example, across our data set
as a whole (encompassing several types of dietary manipulation), dietary stress in parents was
seen to significantly increase pro-inflammatory cytokine levels measured in offspring (overall
d = 0.575). All studies included in our analysis were from experiments in which the offspring
were raised on a normal or control diet, so our findings suggest that a nutrition-dependent
immune state can be inherited, and that this immune state is maintained in the short term,
despite offspring returning to an ‘optimal’ diet. We demonstrate how the geometric
framework for nutrition can be used to disentangle the role that different forms of dietary
manipulation can have on intergenerational immunity. For example, offspring B-cell
responses were significantly decreased when parents were raised on a range of different diets.
Similarly, our approach allowed us to show that a parental diet elevated in protein (regardless
of energy composition and relative to a control diet) can increase expression of inflammatory
markers while decreasing B-cell-associated markers. By conducting a systematic review of
the literature, we have identified important gaps that impair our understanding of the
intergenerational effects of diet, such as a paucity of experimental studies involving increased
protein and decreased energy, and a lack of studies directed at the whole-organism
2
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
consequences of these processes, such as immune resilience to infection. The results of our
analyses inform our understanding of the effects of diet on physiological state across diverse
biological fields, including biomedical sciences, maintenance of agricultural breed stock and
conservation breeding programs, among others.
Key words: cytokines, diet composition, inflammation, macronutrients, nutritional geometry,
obesity.
CONTENT
SI. Introduction.........................................................................................................................4II. Methods..............................................................................................................................7
(1) Literature search..............................................................................................................7(2) Inclusion and exclusion criteria......................................................................................7(3) Data extraction................................................................................................................8
(a) Dietary manipulation measurements and the geometric framework for nutrition......9(b) Immunological trait measurements...........................................................................11
(4) Effect-size calculation...................................................................................................11(5) Missing data..................................................................................................................12(6) Meta-analysis and meta-regression models..................................................................13(7) Publication bias.............................................................................................................14
III. Results...........................................................................................................................15(1) Available data................................................................................................................15(2) Overall effects...............................................................................................................17(3) Heterogeneity................................................................................................................18(4) Effects of diet type........................................................................................................18(5) Publication bias.............................................................................................................20
IV. Discussion.....................................................................................................................20(1) Immune complexity: parental diet differentially affects aspects of offspring immunity
21(2) Dietary complexity: different parental dietary compositions have different effects on offspring immunity...............................................................................................................23(3) Manipulating parental nutrition can lead to an inflammatory state in offspring...........25(4) Whole-organism consequences of parental dietary manipulation................................27
V. Conclusions.......................................................................................................................28VI. Acknowledgements.......................................................................................................29VII. References.....................................................................Error! Bookmark not defined.VIII. Supporting information
3
50
51
52
53
54
55
56
5758
5960616263646566676869707172737475767778798081828384858687
I. INTRODUCTION
It has been widely shown that plants and animals transmit information to their offspring non-
genetically as a function of prevailing environmental conditions, and that these processes are
an important component of inter-individual variation in phenotype. In particular, parents may
transmit information about the state of the nutritional (e.g. Godfrey, Gluckman & Hanson,
2010; Hibshman, Hung & Baugh, 2016), pathogenic (Pigeault et al., 2016), physiological
(e.g. Herman & Sultan, 2016), toxicological (e.g. Pölkki, Kangassalo & Rantala, 2012) or
psychological (e.g. Klengel, Dias & Ressler, 2016; Bohacek & Mansuy, 2015) environment.
Transmitted information may increase fitness if offspring phenotype can be modified
accordingly to optimise interaction with the environment. Intergenerational transmission
occurs if parents provide information directly to gametes, embryos or offspring, via a variety
of physiological mechanisms, such as the transmission of antibodies or hormones (Skinner,
2008). Similarly, ‘transgenerational’ transmission occurs via the germline (i.e. without direct
exposure of the offspring to the stimulant) through epigenetic modifications (Skinner, 2008).
Immune priming is the provision of immune profiles, or other information about the
pathogenic environment, from parent to offspring (Pigeault et al., 2016; Poulin & Thomas,
2008). Female parents typically have greater opportunity to provide immune priming to
offspring than male parents, as the former have greater opportunity for provisioning in
general than the latter [e.g. via direct transmission of antibodies (Grindstaff, Brodie &
Ketterson, 2003)], although paternal effects have been detected (e.g. Keightley, Wong &
Lieschke, 2013; Sternberg, de Roode & Hunter, 2015). Paternal and maternal effects may
even have differential outcomes for offspring phenotype (e.g. Triggs & Knell, 2012).
One environmental mechanism that may modulate the relationship between parental and
offspring immunity is diet. There is a close relationship between diet and immunity across the
4
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
animal kingdom: poor nutrition early in life can negatively affect many adult traits, including
disease resistance (Gebhardt & Newberne, 1974; Valtonen et al., 2012). The reasons for this
relationship are complex and may be attributed to the cost of investing in immunity, so that
immune function must be traded off against investment in other traits (Lochmiller &
Deerenberg, 2000). For example, malnutrition and/or nutrient deficiency may directly result
in parents provisioning fewer resources to offspring, which then themselves have fewer
resources available to invest in immune development, although indirect mechanisms via
imprinting may also occur, especially under paternal effects (Triggs & Knell, 2012).
Conversely, poor nutritional state may be indicative of high population density, which brings
increased disease risk. As a consequence, evolution may favour increased immune function in
offspring during periods of under-nutrition (Shikano et al., 2015). In addition to diet itself
providing the signal for intergenerational immune priming, variation in immune receptivity
may also be generated indirectly via the effects of diet on the microbiota, which can in turn
also be transmitted from parent to offspring (Berghof, Parmentier & Lammers, 2013; Knorr
et al., 2015). Taken together, evidence suggests that the nutritional state of parents may be
expected to have complex effects on the immunological state of their progeny.
A synthetic understanding of the impacts of parental nutrition on offspring immunity would
inform several fields of enquiry. Environmental fluctuations may mean that the conditions
experienced by offspring correlate poorly with those experienced by their parents. If
intergenerational effects occur, environmental change could result in a mismatch between the
supposedly information-rich cues a parent passes to their offspring’s phenotypic
development, and the truly ‘ideal’ phenotype for prevailing conditions – a maladaptive
intergenerational effect (Godfrey et al., 2007). In biomedical sciences, intergenerational
effects based on a mother’s nutritional state may be responsible for many negative health
outcomes, including increased risk of obesity and metabolic disease risk (Bruce & Hanson,
5
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
2010; Godfrey et al., 2007). In conservation biology, non-genetic intergenerational processes
could impact the success of breed-for-release programs: cues that breeders receive from
captive environments may lead to production of offspring with phenotypic traits that are
poorly suited to the wild, which may then show poor survival when released (e.g. Evans et
al., 2014; Grueber et al., 2017; Williams & Hoffman, 2009). In agriculture, feed provided to
breeders may impact the immune capacity of offspring through intergenerational effects, with
potentially massive economic costs if immunity is impaired (e.g. Berghof et al., 2013).
Literature on the effects of nutrition on intergenerational immune priming is diverse, with
experimental data originating from fields as wide-ranging as evolutionary ecology,
parasitology, biomedical research and agriculture. Similarly, effects of parental diet on a
broad diversity of offspring phenotypes associated with immunity have been examined,
including immune gene expression (e.g. Iwasa et al., 2015), immune profiles (e.g.
Tuchscherer et al., 2012), pathogen susceptibility (e.g. Kangassalo et al., 2015) and
survivorship (e.g. Enting et al., 2007). In this study, we statistically collate the findings of
published experimental studies on the topic of intergenerational effects of diet on immunity.
We use a systematic review of the literature coupled with meta-analytic methods to test for
overall trends and inform our understanding of these fundamental processes. Our approach
enables us to distinguish the degree to which different results among studies can be attributed
to sampling variation, rather than true underlying differences (Senior et al., 2016a). Using
quantitative techniques we aim to identify those offspring immunological traits that are most
responsive to different types of nutritional manipulation in parents.
6
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
II. METHODS
(1) Literature search
Our systematic review and meta-analyses target experimental studies that impart
macronutrient stress on an individual (F0), before quantifying immune function in individuals
of the subsequent generation they produced (F1). We performed a systematic literature search,
following established preferred reporting items for systematic reviews and meta-analysis
(PRISMA) protocols as closely as is possible for a multi-species study (Nakagawa & Poulin,
2012; Shamseer et al., 2015). The online search was conducted on 7 June 2016 in both the
ISI Web of Science (“topic” search, encompassing title, abstract, author keyword and
Keyword Plus®) and Scopus (encompassing title, abstract, keyword) databases using the
following keyword combinations: ‘(matern* OR parent*) AND (nutrition* OR diet* OR
calorie* OR food) AND (epigen* OR intergeneration* OR transgeneration* OR offspring)
AND (immun* OR infect* OR parasite*) NOT human NOT cancer’. These online searches
yielded 1,764 and 695 records, respectively. Duplicates were removed, and titles, abstracts
and full texts screened for suitability (Fig. 1).
(2) Inclusion and exclusion criteria
Studies were included in our analysis only if they satisfied the following criteria: (i) the
research must have been conducted under an experimental setting (e.g. excluding
observational field data) with a control parental group, which was fed a ‘normal’ or ‘habitual’
diet (as defined by the authors); (ii) the experimental treatment must have been a quantifiable
macronutrient manipulation, or calorie restriction, relative to controls (see Section II.3a), in
the parental generation (F0); (iii) the study must report quantifiable measures of immune
function correlates (including candidate gene expression analysis and related measures;
Section II.3b) in the first-generation offspring (F1), with or without an immune challenge,
7
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
excluding generalised (non-specific) measures of the inflammatory response; (iv)
experiments were conducted using non-mutant strains, and animals were not subject to any
other interventions.
In addition to those study-design considerations listed above, studies were also excluded if
they did not report the mean of their measured trait and sample size of F0 and F1 groups
(missing trait variances were imputed, Section II.5), or if measurements for F1 offspring
could not be unambiguously tracked back to parental treatment schemes. Available data from
all non-human animal species were included in our analysis. In all, 34 articles included all the
data for effect-size calculation in the main data set, although we were able to include an
additional four studies using multiple imputation of missing data (see Section II.5) (Table 1),
giving a total of 38 articles included.
(3) Data extraction
For each study that met our eligibility criteria, we extracted the mean and, where possible,
standard deviation (SD; calculated from associated measures where necessary) of a measured
immune trait in the offspring of parents that had been subjected to experimental
macronutrient intake manipulation, as well as corresponding control data. The number of
offspring on which each measurement was based was also extracted. We extracted data from
tables presented in the main text of each study, in supplementary material, or from figures
using GraphClick (http://www.arizona-software.ch/graphclick/). For each study and effect
size we also recorded data on covariance structure (e.g. multiple contrasts with the same
control group data) and moderator variables (i.e. factors that may affect the results), namely
study ID, species, and the type of dietary manipulation.
8
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
(a) Dietary manipulation measurements and the geometric framework for nutrition
Here, nutritional stress is examined using the geometric framework (GF) for nutrition
(Raubenheimer & Simpson, 1993; Simpson & Raubenheimer, 2012). The GF allows diet
compositions to be explicitly defined via nutritional geometry wherein food nutrient
components are expressed as axes in a Cartesian coordinate space (a ‘nutrient space’). Recent
application of the GF to immunological studies has demonstrated that an animal’s nutritional
state is central to their immune function (reviewed by Simpson et al., 2015). Our study is the
first, however, to investigate the impacts of intergenerational nutrient stress using the GF.
In this study, we characterise experimental diets according to their macronutrient composition
in a nutrient space delimited by two axes, with the percentage of dietary energy derived from
protein expressed on the x-axis, and non-protein-derived energy expressed on the y-axis (Fig.
2). We consider the habitual food on which the control group was fed in each study as the
optimal food (i.e. the black food rail in Fig. 2) and consider experimental diets which deviate
geometrically from this optimum as stimulators of macronutrient stress. Employing this
nutrient space allows us to categorise nutritional stress qualitatively (defined as a
manipulation of the ratio of protein to non-protein calories in the food; grey rails in Fig. 2),
and quantitatively (whether access to total energy was manipulated relative to the optimal
diet; position along the black or grey rails in Fig. 2) (following Raubenheimer et al., 2016). A
variety of physiological and life-history traits are known to be influenced by both dietary
macronutrient composition (Le Couteur et al., 2016; Machovsky-Capuska et al., 2016) and
dietary restriction (Robertson & Mitchell, 2013; Senior et al., 2015). In our analysis, all
parental macronutrient intake treatment and control groups were defined according the
following criteria: (a) was there an increase or decrease in total energy for the treatment diet,
relative to control, or was the total energy offered by each diet the same? (b) Was there an
increase or decrease in the percentage composition of protein for the treatment diet, relative
9
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
to control, or was the percentage protein of each diet the same? (c) Were the treatment
animals maintained under an ‘intake restriction’ protocol? The latter could be achieved by
either restricting the total food intake of the treatment animals, or by feeding the treatment
animals an equivalent amount (per unit intake) of food to the control, however with the
concentration of the treatment animals’ food being diluted with a non-nutritive bulking agent.
For studies that did not explicitly report the percentage contribution to total energy made by
each macronutrient in their treatment and control-group diets, we calculated these values
using the diet recipes provided by authors in the main text. Some recipes utilised complex
food ingredients (e.g. yeast), each of which contributed to total diet energy via multiple
macronutrients simultaneously. In these instances, we used the Food Standards Australia
New Zealand NUTTAB food database
(http://www.foodstandards.gov.au/science/monitoringnutrients/nutrientables/pages/
default.aspx) to determine the macronutrient percentage composition of each complex food.
We then summed the energy contribution made by each macronutrient across all the
ingredients in an individual recipe to determine the total contribution each macronutrient
made to total dietary energy. In doing so, we followed the standard rule that protein and
carbohydrate each contribute 4 kcal g–1 and fat contributes 9 kcal g–1 (following Senior et al.,
2016b). In some cases, authors reported using proprietary diets. In these cases, we extracted
the energy density and the protein composition of these diets from the proprietors’ product
nutritional information labels or product information sheets available online. All data were
extracted from studies wherein the offspring of parental treatment and control groups were all
raised on the control diet.
10
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
(b) Immunological trait measurements
A wide diversity of immunological parameters, proteins, tissues, immunogenetic expression
and other traits were examined in the literature we extracted. To facilitate our analysis of the
general effects of parental diet on offspring immunological traits, traits were broadly
classified into nine categories as follows: (1) adaptive, B-cell associated: B cells and
associated proteins/genes including immunoglobulins; (2) adaptive, T-cell associated: T cells
and associated cells/proteins including cytotoxic T cells, T helper cells and Th1 and Th2
cytokines; (3) innate cells: all innate cell counts or markers of these including neutrophils,
macrophages and eosinophils; (4) pro-inflammatory cytokines: all pro-inflammatory
cytokines; (5) inflammation associated, other: other genes/proteins associated with
inflammation and the innate immune response including inflammasome components, Toll-
like receptor (TLR), complement and anti-bacterial components; (6)
anti-inflammatory/immune regulation: cells and cytokines associated with anti-inflammation
or down-regulation of immune response; (7) organ mass: mass of immune organ including
thymus and spleen; (8) resistance to infection: experimental data from studies that infected or
otherwise challenged subjects and recorded their resistance to the disease either in the form of
percentage mortality, resistance to sublethal infection or delayed development; (9)
leukocytes, other: leukocyte or lymphocyte counts that cannot be assigned to one of the above
categories due to non-specificity of counts.
(4) Effect-size calculation
Our aim is compare the difference in mean values of immunological traits of offspring whose
parents have been exposed to different dietary treatments. Because we are combining data
from a range of different trait measurements and species it is necessary to convert the
outcomes of each experiment to a common scale. In meta-analysis, this is done using a
11
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
standardised effect size (Nakagawa et al., 2017; Rosenberg, Rothstein & Gurevitch, 2013).
We used the effect size Hedge’s d (hereafter d), which is one of the most widely used effect
sizes for quantifying the difference in population means of two treatment groups in an
experiment or clinical trial (Grissom & Kim, 2001; Gurevitch, Curtis & Jones, 2001; Hedges,
1981). For each experimental (manipulated diet) and control-group combination in a study,
we calculated d following equation 14 in Nakagawa & Cuthill (2007). Measures of d are
positive if the mean measurement of the trait of interest is greater in the experimental group
than in the corresponding control group, and vice versa.
In meta-analysis, the variable power of each effect size should be taken into account. Effect-
size estimates from experiments with larger sample sizes should be considered as more
precise than those from studies with low sample sizes. This is done by estimating the
sampling variance (the square of the standard error) associated with each effect size. We
calculated the sampling variance associated with each value of d following equation 17 in
Nakagawa & Cuthill (2007). All calculations and subsequent analyses were performed using
R V 3.4.0 (R Core Team, 2016). Our input data file, along with the associated code for
conducting all analyses, are provided as online Supporting Information, Appendix S1 and S2,
respectively.
(5) Missing data
In a number of cases, experiments did not report the standard deviation associated with a
mean. Rather than completely exclude these data from our analyses, we used multiple
imputation to impute missing standard deviations [full details provided in Appendix S3.1
(Besson et al., 2016; Nakagawa, 2015)]. Multiple imputation was performed using the “mice”
function in the package mice (van Buuren & Groothuis-Oudshoorn, 2011). To explore the
sensitivity of our conclusions to the use of multiple imputation, we also repeated our analyses
12
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
after dropping all data with missing standard deviations. The results of these less-inclusive
analyses are presented in Fig. S2 and Table S1, and provide the same qualitative conclusions
as the results presented in the main text.
(6) Meta-analysis and meta-regression models
Effect sizes from traits in each immunological response category (Section II.3) were analysed
separately. Quantitative analyses were only applied to those data subsets with greater than
five effect sizes (Fig. 1). For each trait category we implemented three models, using the
‘rma.mv’ function in the package metafor (Viechtbauer, 2010) for R.
The first model was a multi-level meta-analysis (MLMA; Nakagawa et al., 2017). This type
of model has two aims: (i) estimate the overall average effect based on all effect sizes within
a trait category, and associated statistical significance, and (ii) estimate the amount of
variation among the effects reported by different studies, termed ‘heterogeneity’ (Higgins &
Thompson, 2002; Senior et al., 2016a). For estimated effects, we interpret effect magnitude
of estimates using the established benchmarks of 0.2, 0.5 and 0.8 as small, medium and large,
and interpret statistical significance using 95% confidence intervals (CI; CIs not spanning
zero are considered statistically significant at α = 0.05 [Nakagawa & Cuthill, 2007]).
Heterogeneity (i.e. variation among studies) was calculated as I2Total (Higgins & Thompson,
2002; Nakagawa & Santos, 2012). I2Total is interpretable as the percentage of variance among
effects that cannot be attributed to sampling. Following widely used (but initially tentatively
proposed) benchmarks we consider values greater than 25%, 50% and 75% as small, medium
and large heterogeneity, respectively [although for a multi-species analysis such as ours it is
not uncommon for I2Total to be greater than 90% (Borenstein et al., 2009; Higgins et al., 2003;
Nakagawa et al., 2017; Senior et al., 2016a)]. In several cases, multiple effect sizes were
13
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
obtained from the same study, so total heterogeneity (I2Total) was partitioned into between-
(I2Study) and within-study (I2
Residual) heterogeneity following Nakagawa & Santos (2012).
After fitting MLMAs for each trait, we went on to implement two random-effects meta-
regressions (REMRs) for each trait. The aim of REMR is to explore the degree to which
variance among effect sizes (i.e. heterogeneity) could be explained by the factors, termed
moderators, that are known to differ among studies (Nakagawa et al., 2017). Here we
explored whether different types of parental dietary manipulation (Section II.3a) may have
different effects on offspring immunological traits. Two REMRs were implemented for each
data subset, exploring the effects of (i) % energy from protein in the diet, and (ii) energy
content of the diet. Moderators were coded as three-level categorical predictors; ‘decrease’,
‘same’ or ‘increase’ in the experimental diet relative to control. Multiple moderators were not
included in the same model simultaneously owing to the relatively small sample sizes in data
subsets.
Like most statistical models, meta-analytic models assume that data (i.e. effect sizes) are
independent (Noble et al., 2017). Where this assumption is violated, and non-independence is
not appropriately addressed, there is an increased risk of Type I error (Noble et al., 2017). We
identified a number of sources of non-independence among our effect sizes, and dealt with
these issues using recommended procedures (detailed in Appendix S3.2).
(7) Publication bias
The results of meta-analyses can be biased if the experimental results that make it into the
published literature on a topic systematically differ from the results of the sum total of
experiments that have been performed on that topic (Rothstein, Sutton & Borenstein, 2005).
Most commonly, it is asserted that non-significant results have a decreased likelihood of
publication (for a range of reasons), meaning that meta-analyses based on the published
14
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
literature lack studies that observe small effects and/or have small sample sizes (Jennions et
al., 2013; Nakagawa et al., 2017). Consequently, where present, publication bias can lead to
upwardly biased meta-analytic estimates. A number of tests have been devised to detect, and
test the sensitivity of results to, publication bias (Møller & Jennions, 2001). Here, we used a
regression-based method (‘Egger’s regression’) to detect statistically significant evidence for
the presence of publication bias in each data subset (i.e. for each immunological trait) (Egger
et al., 1997; Nakagawa & Santos, 2012). For data sets where we detect evidence of
publication bias using Egger’s regression, we used ‘trim-and-fill analysis’ to evaluate the
sensitivity of our results to publication bias (Duval, 2005). Trim-and-fill analyses estimate the
number of ‘missing’ (i.e. unpublished) effect sizes, and the degree to which the results of
meta-analysis might be biased by these missing results. Full details of the implementation of
these tests can be found in Appendix S3.3.
III. RESULTS
(1) Available data
After filtering, and consolidation of multiple related measures of immune capacity from the
same group, 38 publications yielded a total of 226 effect sizes (71 of which were based on
imputed standard deviations), across our nine categories of immune function. Only one
category, Leukocytes (Other), had fewer than five effect sizes (Fig. 1), and was therefore
excluded from subsequent analysis. The mean sample size of treatment groups that
contributed to each effect size in our data set was 12.2 individuals (SD = 20.2, range 4 – 211),
although note that the same group of animals may contribute to multiple effect sizes (for
example, if multiple broad categories of immunological measurements are made from the
same individuals).
15
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
Dietary manipulations were typically applied to mothers (N = 181 effect sizes, 80.1%) with
other manipulations applied to both parents (N = 45, 19.9%). None of the studies included in
our analyses manipulated the diets of fathers alone. Over half of our data (N = 170 effect
sizes, 75.2%) came from analyses where both the treatment and control groups were fed ad
libitum. For the remainder, 36 effect sizes (15.9%) were from studies where treatment diets
were fixed and controls ad libitum, and 20 (8.8%) were from studies where diets of both
groups were fixed. Diet treatments varied widely, with almost all combinations of
‘decreased’, ‘same’, and ‘increased’ energy and protein observed, although three diet types
dominated the data set: decreased energy and protein, decreased energy with same protein,
same energy with decreased protein; together comprising 67.3% of the data set (Table 2).
Our data set incorporated findings collected from studies of 13 species, primarily studies of
rodents (Norway rat Rattus norvegicus, mouse Mus musculus, greater long-tailed hamster
Tscherskia triton), but also including three species of Artiodactyla (goat Capra aegagrus
hircus, boar Sus scrofa, sheep Ovis aries), a primate (macaque Macaca fuscata), two
Galliformes birds (chicken Gallus gallus, Japanese quail Coturnix japonica), three
lepidopteran moth species (Indian mealmoth Plodia interpunctella, cabbage looper
Trichoplusia ni and greater wax moth Galleria mellonella) and a dipteran mosquito
(Anopheles gambiae) (Table 1). Although there are important differences in immunity across
major animal taxa (such as invertebrates versus vertebrates), there are many crucial
similarities too (Buchmann, 2014). Our extracted data set was distributed such that eight of
our nine traits were represented by multiple species (maximum six species contributing data
to a given trait), and nine of the 13 study species were represented by data across more than
one trait (see Appendix S1). Studies of invertebrates contributed only to two traits:
inflammation associated (other) and resistance to infection. Studies of vertebrates contributed
16
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
data to all immunological categories; R. norvegicus was the only species represented across
all traits.
(2) Overall effects
We first report on the main results of our meta-analysis, indicating global trends across the
data set and levels of heterogeneity (i.e. variability across studies). Overall, we observed
intergenerational impacts of parental nutrition on a number of offspring immunological
processes. Considering our data set as a whole, we observed that dietary manipulation in
parents led to a small, but non-statistically significant decrease in adaptive, B-cell-associated
traits (MLMA d = 0.205, 95% CI = –0.444 – 0.035; N = 37; Fig. 3A). We detected no overall
effect of parental diet on offspring adaptive T-cell-associated traits or innate cells (Fig. 3B
and 3C). Overall, parental dietary manipulations lead to increased pro-inflammatory
cytokines in the offspring, and this effect was moderate and statistically significant (MLMA
d = 0.575, 95% CI = 0.166 – 0.983; N = 55; Fig. 3D). Similarly, we detected a moderate and
statistically significant effect of parental diet on other inflammation-associated traits (MLMA
d = 0.447, 95% CI = 0.054 – 0.840; N = 32; Fig. 3E). There was a small negative effect of
parental diet on anti-inflammatory and immune regulatory offspring traits, although the
sample size for this effect was small and the associated CI was wide and spanned zero
(MLMA d = –0.324, 95% CI = –0.890 – 0.242; N = 15; Fig. 3F). There was no effect of
parental dietary manipulations on offspring immune organ mass (Fig. 3G). Finally, offspring
resistance to infection was estimated to be slightly depressed by parental dietary
manipulations, however the effect was non-significant (MLMA d = –0.178, 95% CI = –0.461
– 0.104; N = 23; Fig. 3F).
17
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
(3) Heterogeneity
Heterogeneity was present in all analyses; its magnitude ranged from low to very high across
our immunological traits (I2total = 29.5 – 98.7%; Table 3), indicating a degree of variation in
the reported effects across different studies. Heterogeneity may result from methodological
differences across studies, and/or species-specific effects. For five traits (adaptive B-cell
associated, adaptive T-cell associated, innate cells, pro-inflammatory cytokines, anti-
inflammatory/immune regulatory), heterogeneity could be attributed completely to variation
at the study level, with I2residual (i.e. within-study variation) near zero (Table 3). That is, all
variation within studies could be attributed to sampling variance, and all variation among
effect sizes, aside from sampling variance, could be attributed to study-specific factors (e.g.
differences in study design and/or species; Table 3). For other inflammation-associated traits,
heterogeneity among effects was attributable to both among-study differences, and within-
study variation (Table 3). For two traits (organ mass and resistance to infection) inter-study
heterogeneity, I2studies, was close to zero, leaving large to moderate amounts of variance
unaccounted for (I2residual ~ 70%; Table 3).
(4) Effects of diet type
We used REMR to evaluate the impacts of particular parental dietary manipulations (protein
and energy content) on offspring immunological traits. In offspring adaptive, B-cell-
associated traits, manipulations that altered protein content isocalorically led to significant,
small decreases, while studies that reduced energy content had little effect (Fig. 3A, see
Tables S3 and S4 for full REMR results unless otherwise stated). A single study measured B-
cell-associated traits in response to a manipulated diet that held percentage of energy from
protein constant, and observed a very large negative effect (Fig. 3A). For adaptive, T-cell-
associated traits, both increasing energy content and manipulating percentage dietary protein
18
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
resulted in slight decreases, but these effects were non-significant, and we detected no
significant differences between type of dietary manipulation for these traits (Fig. 3B). For
innate cells, the estimated effects of all dietary manipulations were associated with poor
precision and were non-significant (Fig. 3C).
Offspring pro-inflammatory cytokines were increased as a result of all forms of parental
dietary stress (Fig. 3D). These effects were particularly large, and statistically significant, for
increases in energy content (Fig. 3D). Other offspring inflammatory traits were also markedly
affected by different forms of parental dietary stress (Fig. 3E). In particular, increasing
dietary energy caused large increases in inflammation, an effect that was significantly greater
in magnitude than for decreases in energy (REMR dincrease–decrease = 0.810, CI = 0.111 – 1.509;
Fig. 3E). Increased protein also had a significantly greater effect on non-cytokine
inflammatory traits than the effect of decreased protein (REMR dincrease–decrease = 0.990, CI =
0.037 – 1.943; Fig. 3E). Offspring anti-inflammatory and immune regulatory traits showed a
statistically significant decrease in response to increased parental dietary energy, the
magnitude of which was greater than that seen for decreased parental dietary energy
(REMR dincrease–decrease = –1.210, CI = –2.136 – –0.285; Fig. 3F), although the latter was based
on just two studies. Both increasing and decreasing protein content of the diet was estimated
to slightly decrease offspring anti-inflammatory traits, although estimates were non-
significant (Fig. 3F).
All forms of dietary stress in the parental generation had negligible mean effects on offspring
organs mass (Fig. 3G). Finally, all studies that measured resistance to infection used a
treatment that decreased energy content, and there were no significant differences between
types of protein manipulation for this trait (Fig. 3H).
19
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
(5) Publication bias
Using Egger’s regression, we found statistically significant evidence for publication bias for
five of our eight analysed traits (see Table S5 for full publication bias results). Trim-and-fill
analyses applied to these data subsets suggested only one missing effect for each of adaptive
T-cell-associated traits, anti-inflammatory/immune regulation, and resistance to infection. For
anti-inflammatory traits and resistance to infection, inclusion of the missing data pulls the
overall effect towards zero, but does not affect the qualitative conclusion drawn. For innate
cells, trim-and-fill analysis indicated two effects were missing, although again, adjustment for
these data does not qualitatively alter the interpretation of the overall effect. Finally, trim-
and-fill analysis suggested five effect sizes missing for our analysis of pro-inflammatory
cytokines, and that this overall result may be upwardly biased by 0.109 (Table S5); after
adjustment for this bias the overall estimated effect for cytokine expression remained
moderate in magnitude (overall d = 0.466).
IV. DISCUSSION
In this study, we systematically reviewed the effects of experimentally induced parental
nutritional stress on offspring immunity. We observed intergenerational impacts of parental
nutrition on a number of offspring immunological processes. In particular, we observed that
pro-inflammatory cytokines and other inflammatory markers were elevated, while anti-
inflammatory markers were decreased. These results point to a systematic intergenerational
effect of dietary manipulation in parents, on the inflammatory state of offspring. We also
found intergenerational effects of dietary manipulation on B-cell-associated processes, but
little to no effect on T-cell-associated responses, gross immune organ masses, nor innate cell
counts. Our meta-analysis also found a negative trend of parental dietary manipulation on
offspring resistance to infection, although the latter pattern was not statistically significant.
20
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
Finally, we were able to identify particular dietary stressors that have a greater or lesser effect
on offspring immunity. For example, increased energy in the parental diet resulted in a large
increase in offspring inflammatory markers, relative to little or no effect when energy was
held constant or decreased (Fig. 3E).
Overall, our results allow us to draw several important inferences about intergenerational
impacts of parental nutrition on offspring immunity and disease resilience, as well as
identifying fruitful areas for future research. Diet and immunity represent two of the most
fundamental mechanisms by which organisms interact with their environment. But both diet
and immunity are highly complex, multi-dimensional traits, leading to difficulties in
untangling the relationships between alternate life-history and physiological strategies, and
predicting responses to perturbations (Ponton et al., 2011).
(1) Immune complexity: parental diet differentially affects aspects of offspring
immunity
Our results show that offspring of parents maintained on a range of sub-optimal diets
displayed down-regulation of their adaptive, B-cell-associated responses (Fig. 3A). By
contrast, parents maintained on sub-optimal diets showed no effect on adaptive, T-cell-
associated responses relative to offspring from parents whose diets were optimal (Fig. 3B).
For juvenile herbivores and omnivores, a compromised nutritional state corresponds to
differential performance in B-cell- and T-cell-mediated immune responses: short- to medium-
term protein and/or energy restriction can inhibit antibody production in juvenile animals,
while T-cell activity, especially that of helper and killer T cells, remains either unchanged or
enhanced (Esquifino et al., 2007; Fernandes, 2008; Good & Lorenz, 1992). Our meta-
analysis suggests that offspring, too, could inherit this nutrition-dependent immune state
(differential regulation of B-cell versus T-cell responses) from nutritionally compromised
21
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
parents, without directly experiencing nutritional stress themselves. The consequences of this
finding are far-reaching, for example in conservation breeding programs that produce animals
for release to the wild. Released animals with reduced antibody production and/or activity
could suffer increased infection risk and decreased antigen recognition. In cases where the T-
cell response is increased, this situation could present an increased risk of developing auto-
immune diseases (Ma et al., 2008; Manzel et al., 2014) and allergies (Good & Lorenz, 1992;
Ho, 2010), and promote inflammation (Cohen, Danzaki & MacIver, 2017; Magrone & Jirillo,
2015).
Exactly how parental macronutrient and/or caloric intake manipulation influences offspring
lymphocyte activity, and potential disease risk, warrants further investigation. In the current
study, the published literature we identified contained a diversity of immunological
measurements, which necessarily required creation of broad B-cell and T-cell response
categories for meta-analysis. However, T-cell type and function vary widely (Magrone &
Jirillo, 2015), and future work should test how different components of T-lymphocyte-
mediated immunity is transmitted in response to variable nutritional states, to provide a more
precise picture of the underlying mechanisms that drive the patterns we report. Furthermore,
T and B lymphocytes mature at different developmental stages in animals (Fadel & Sarzotti,
2000; Garcia et al., 2000), and future studies should aim to differentiate the responses of
directly versus indirectly inherited immunoglobulins. The current analysis included B-cell
responses from diverse animal species, including birds and mammals who inherit maternal
antibodies through yolk or milk. It is possible that these broad groupings obscure some of the
subtleties of offspring nutrient-dependent responses. Our need to create these broad
immunological response-variable groups for meta-analysis is itself noteworthy, and
demonstrates a paucity of current work investigating the intergenerational immunological
22
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
impact of poor nutrition, especially with specific immunological responses (Manzel et al.,
2014).
Several categories did not show any significant change in response to dietary manipulations,
including innate cell counts. As pro-inflammatory cytokines were increased we may expect
the counts of cells producing these cytokines, such as macrophages, to increase also.
However, studies that focussed on ‘innate cells’ were diverse, incorporating many measures
of diverse cellular and biochemical markers, each of which may respond differently to dietary
manipulations. Examining each of these independently would have reduced our ability to test
for general patterns. In addition, offspring organ mass, which included both thymus and
spleen masses, was not significantly affected by parental dietary manipulation. While it is
possible that these organs are unaffected by dietary effects, we observed a high degree of
unexplained heterogeneity for these traits, i.e. variation among studies that cannot be
explained by dietary manipulation. Thus, the lack of a statistically significant change may
also be due to differential responses of spleen and thymus, and/or among-study variation in
the developmental period at which masses were measured. Developmental stage is
particularly important for the thymus, which grows then atrophies during development
(Shanley et al., 2009).
(2) Dietary complexity: different parental dietary compositions have different effects on
offspring immunity
In this meta-analysis, we used the geometric framework (GF) to avoid confounding changes
in diet quantity with changes in diet nutrient proportions (Ponton et al., 2011; Simpson &
Raubenheimer, 2012). For example, 38 effect sizes in our data set were from ‘high fat’
dietary manipulations (Table 2), but these were achieved via four different types of
macronutrient and energy manipulations (i.e. the absolute compositions of ‘high fat’ diets
23
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
varied considerably). Similarly, over half of the effect sizes included in our study were
manipulations of the protein content of the diet, but only 15.9% of the data set were from
experimental increases in protein (regardless of dietary energy content) (Table 2).
In humans, high-protein diets may assist in weight loss and controlling obesity, due to the
appetite-suppressant effect of protein (Astrup, 2005; Gosby et al., 2014; Simpson &
Raubenheimer, 2005; Weigle et al., 2005). While the health benefits of reducing excess
adipose tissue in animals, including humans, are well understood (Han, Sattar & Lean, 2006;
van der Kooy et al., 1993; Zimmet, Alberti & Shaw, 2001) recent studies have shown that a
high-protein diet, with concurrent reduction in total energy from carbohydrates and fat,
causes health and lifespan attenuation in many species (Hunt et al., 2004; Lee et al., 2008;
Solon-Biet et al., 2014). Our meta-analysis indicates that increasing the percentage of protein
beyond a species’ optimum in parental diets (regardless of energy intake) can significantly
increase inflammatory biomarkers in offspring (Fig. 3D, 3E) and decrease B-cell-associated
markers (Fig. 3A). Thus, increasing parental dietary protein could have important effects on
offspring immunity. Nutrition interacts with immunity within the lifespan of an individual by
influencing disease tolerance, immune activation following pathogen exposure, and longevity
following infection (Lazzaro & Little, 2009; Ponton et al., 2013). Work on larval insects has
shown that increasing the relative proportions of different macronutrients can variously
optimise different immune traits, with no particular protein to carbohydrate proportion able to
optimise all immune components within a species (reviewed in Ponton et al., 2013). Given
the complexity of disease risk in nature (Lazzaro & Little, 2009), our finding that a nutrition-
dependent immune state can be inherited, and that this immune state is maintained in the
short-term despite offspring returning to an ‘optimal’ diet, is important in many fields of
enquiry. For example in conservation, long-term captive diets may perturb offspring immune
states in a manner not seen in the wild. Obesity is a growing problem for captive animals (e.g.
24
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
Clubb et al., 2008; Flacke et al., 2016; Rawski & Józefiak, 2014; Rose & Roffe, 2013), and
macronutrient manipulations are beginning to be used for weight control in captive animals
(Aoki et al., 2015; Simpson & Raubenheimer, 2012). It is unclear how these dietary
manipulations may impact the survival of offspring released to the wild, to supplement
threatened natural populations. Future work will need to establish whether offspring food
choices, if given a choice, can override their inherited immune states.
(3) Manipulating parental nutrition can lead to an inflammatory state in offspring
There is increasing focus on links between nutrition, energy consumption, obesity and
inflammation. In humans, obesity is characterised by chronic over-nutrition and excessive
adiposity. Obesity is co-morbid with chronic inflammation (Das, 2001) where circulating
pro-inflammatory cytokine and protein levels are elevated relative to those of non-obese
individuals (Dandona, Aljada & Bandyopadhyay, 2004; Das, 2001; de Heredia, Gómez-
Martínez & Marcos, 2012; Fantuzzi, 2005). Chronic inflammation is thought to contribute to
serious illnesses, including Alzheimer’s disease, Type II diabetes (Dandona et al., 2004; Das,
2001; de Heredia et al., 2012; Heilbronn & Campbell, 2008), cancer, and cardiovascular
disease (Coussens & Werb, 2002; Deng et al., 2016). It unclear whether chronic
inflammation is driven by diet (Dandona et al., 2004; Deng et al., 2016), lifestyle [i.e. a lack
of exercise (Das, 2001)], or due to excess adipose tissue (Dandona et al., 2004; de Heredia et
al., 2012; Fantuzzi, 2005; Galic, Oakhill & Steinberg, 2010; Heilbronn & Campbell, 2008),
as these conditions typically occur simultaneously. ‘Excess’ adipose tissue is thought to
contribute to chronic inflammation through production of pro-inflammatory proteins by
stressed, enlarged adipocytes (Fantuzzi, 2005; Fernández-Sánchez et al., 2011; Galic et al.,
2010), or through cytokine-releasing macrophages that are harboured at high density within
the adipose of obese people (Arango Duque & Descoteaux, 2014; Fantuzzi, 2005; Galic et
25
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
al., 2010). Thus, compared to lean individuals, obese individuals suffer chronic inflammation
because, proportionally, obese bodies contain more pro-inflammatory tissue. However, our
meta-analysis of non-human experimental data was inconsistent with this explanation: we
found significantly higher levels of pro-inflammatory cytokines and other inflammatory
biomarkers in the offspring of parents subjected to macronutrient stress, relative to control
offspring (Fig. 3D, 3E). Offspring inflammatory cytokines and other inflammatory makers
were particularly increased in experiments that increased the energy content of parental diets,
relative to the offspring of parents on control diets (Fig. 3D, 3E), while anti-inflammatory
markers were significantly decreased in these experiments (Fig. 3F). We note that all
treatment offspring included in our analysis were returned to the species’ optimal or control
diet at birth or hatch. These findings suggest that the pro-inflammatory state representative of
a diet elevated in energy can be inherited by offspring that consume optimal diets. Broadly
speaking, the consequences for offspring with elevated pro-inflammatory biomarkers extend
beyond altered disease risk. Elevation in pro-inflammatory cytokines can impact insulin
signalling and promote insulin resistance across a broad range of cells, including adipocytes,
liver and muscle cells (Lavallard et al., 2012). Increased levels of interleukin cytokines can
alter animal fever responses, and even induce fevers similar to those produced by bacterium
infection (Kozak et al., 1998). Increased cytokine activity has also been linked to impairment
of learning and memory, mood alteration (Donzis & Tronson, 2014) and altered sleep
patterns (Majde & Krueger, 2005). Animal breeding programs that manipulate diets in this
way may thus produce animals that experience greater physiological stress than animals on
non-manipulated diets, negatively impacting the suitability of those animals for their intended
purpose, whether that is releasing animals to supplement dwindling wild populations (e.g.
conservation restoration programs) or maintaining a robust breeding stock (e.g. agriculture).
26
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
(4) Whole-organism consequences of parental dietary manipulation
Despite our data set including many studies examining the intergenerational effects of diet on
physiological and genetic measures of immunity, our search provided few data on the
ultimate adaptive consequences of these processes, such as effects on offspring pathogen
resilience or offspring survival. This is reflected in the small number of experiments (N = 23)
that were available for our category ‘resistance to infection’. We would predict that
immunological perturbations may lead to decreased immune resilience, and this is what we
observed, although the trend was not statistically significant (Fig. 3H). Furthermore, under
our classification only two types of dietary manipulation were used: decreased protein
coupled with decreased energy and same protein with decreased energy. For example, Boots
& Roberts (2012) report experimental data from a study of Indian meal moth Plodia
interpunctella, in which the resource environment of mothers (i.e. the feed) was manipulated
in order to determine the effect on offspring survival after challenge with a virus. The dietary
manipulation comprised the substitution of carbohydrate in the diet with an indigestible
bulking agent. Offspring of mothers experiencing poor nutritional state had improved
resilience to parasites, a finding that is interpreted by the authors as indicating life-history
‘optimisation’, or priming (Boots & Roberts, 2012).
Good experimental models for unifying the physiological and evolutionary mechanisms
underpinning nutrition-mediated intergenerational immune priming are species in which it is
possible to cultivate large sample sizes, and with well-known disease processes. An example
is Lepidoptera, many species of which are important agricultural pests and thus have well-
understood pathogen interactions through research into biocontrol agents [e.g. Plodia
interpunctella and its granulosis virus PiGV (Boots, 2000; Boots & Roberts, 2012);
Trichoplusia ni and the bacteria Bacillus thuringiensis (Ericsson et al., 2009; Shikano et al.,
2015); Plutella xylostella and the endoparasitoid wasp Cotesia plutellae (Ibrahim & Kim,
27
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
2006)]. Studies of the role of nutrition in the interaction between insects and their pathogens
have a long history (e.g. Benz, 1987), and more-recent studies provide a well-defined
nutritional framework for lepidopterans (e.g. Cotter et al., 2011; Povey et al., 2014).
Furthermore, all of the cytokine studies included in our current analysis were undertaken in
mammals, but related processes could also be measured in Lepidoptera (e.g. Ibrahim & Kim,
2006), work that may be facilitated by the pre-existence of reference genomes (Challis et al.,
2016).
V. CONCLUSIONS
(1) The results of our meta-analysis reveal that parental dietary manipulations can impact the
immunity of offspring, including increased pro-inflammatory cytokines and other
inflammatory biomarkers, as well as B-cell responses.
(2) Intergenerational immunological responses occurred even when offspring were raised on
a ‘normal’ or ‘control’ diet, allowing us to infer that a pro-inflammatory state can be inherited
by individuals who are themselves raised on a ‘normal’ diet.
(3) Our use of explicit dietary characterisation via the geometric framework allowed for a
nuanced approach to understanding the intergenerational effects of dietary stress on offspring
immunity, and allowed us to identify combinations of dietary energy and macronutrient
manipulations that are poorly studied in the literature.
(4) Few studies have examined the adaptive consequences of intergenerational effects of diet
on offspring immunity, such as impacts on disease resistance or offspring survival. Studies in
this area will help to understand better the evolutionary mechanisms driving the associations
between dietary stress and the immunological responses we observed.
28
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
VI. ACKNOWLEDGEMENTS
We thank two anonymous reviewers for their comments on this manuscript. This research
was supported by a grant from the University of Sydney Faculty of Veterinary Science (now
Sydney School of Veterinary Science) CPC Intramural Fellowship to C.E.G.; A.M.S. was
supported by a Judith and David Coffey Fellowship from the University of Sydney.
VII. REFERENCES
AOKI, K., MITSUTSUKA, S., YAMAZAKI, A., NAGAI, K., TEZUKA, A. & TSUJI, Y. (2015).
Effects of seasonal changes in dietary energy on body weight of captive Japanese
macaques (Macaca fuscata). Zoo Biology 34, 255–261.
ARANGO DUQUE, G. & DESCOTEAUX, A. (2014). Macrophage cytokines: involvement in
immunity and infectious diseases. Frontiers in Immunology 5, 1–12.
ASTRUP, A. (2005). The satiating power of protein—a key to obesity prevention? The
American Journal of Clinical Nutrition 82, 1–2.
BENZ, G. (1987). Environment. In Epizootiology of insect diseases (ed. J. R. Fuxa and Y.
Tanada), pp. 177–214. John Wiley & Sons, New York.
BERGHOF, T. V. L., PARMENTIER, H. K. & LAMMERS, A. (2013). Transgenerational
epigenetic effects on innate immunity in broilers: an underestimated field to be
explored? Poultry Science 92, 2904–2913.
BESSON, A. A., LAGISZ, M., SENIOR, A. M., HECTOR, K. L. & NAKAGAWA, S. (2016). Effect
of maternal diet on offspring coping styles in rodents: a systematic review and meta-
analysis. Biological Reviews 91, 1065–1080.
BILBO, S. D. & TSANG, V. (2010). Enduring consequences of maternal obesity for brain
inflammation and behavior of offspring. Faseb Journal 24, 2104–2115.
29
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
BLACKMORE, H. L., PIEKARZ, A. V., FERNANDEZ-TWINN, D. S., MERCER, J. R., FIGG, N.,
BENNETT, M. & OZANNE, S. E. (2012). Poor maternal nutrition programmes a pro-
atherosclerotic phenotype in ApoE -/- mice. Clinical Science 123, 251–257.
BOHACEK, J. & MANSUY, I. M. (2015). Molecular insights into transgenerational non-genetic
inheritance of acquired behaviours. Nature Reviews Genetics 16, 641–652.
BOOTS, M. (2000). Density-independent resource limitation and the transmission of an insect
pathogen. Oecologia 124, 172–175.
BOOTS, M. & ROBERTS, K. E. (2012). Maternal effects in disease resistance: poor maternal
environment increases offspring resistance to an insect virus. Proceedings of the
Royal Society B: Biological Sciences 279, 4009–4014.
BORENSTEIN, M., HEDGES, L. V., HIGGINS, J. P. T. & ROTHSTEIN, H. R. (2009). Introduction
to meta-analysis. John Wiley & Sons, Ltd, Chichester.
BRUCE, K. D. & HANSON, M. A. (2010). The developmental origins, mechanisms, and
implications of metabolic syndrome. The Journal of Nutrition 140, 648–652.
BUCHMANN, K. (2014). Evolution of innate immunity: clues from invertebrates via fish to
mammals. Frontiers in Immunology 5, 459.
CALDER, P. C. & YAQOOB, P. (2000). The level of protein and type of fat in the diet of
pregnant rats both affect lymphocyte function in the offspring. Nutrition Research 20,
995–1005.
CHADIO, S., KATSAFADOU, A., KOTSAMPASI, B., MICHAILIDIS, G., MOUNTZOURIS, K. C.,
KALOGIANNIS, D. & CHRISTODOULOU, V. (2016). Effects of maternal undernutrition
during late gestation and/or lactation on colostrum synthesis and immunological
parameters in the offspring. Reproduction Fertility and Development 28, 384–393.
CHALLIS, R. J., KUMAR, S., DASMAHAPATRA, K. K. K., JIGGINS, C. D. & BLAXTER, M.
(2016). Lepbase: the Lepidopteran genome database. bioRxiv, DOI: 10.1101/056994.
30
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
CHEN, J.-H., MARTIN-GRONERT, M. S., TARRY-ADKINS, J. & OZANNE, S. E. (2009).
Maternal protein restriction affects postnatal growth and the expression of key
proteins involved in lifespan regulation in mice. PLoS ONE 4, e4950.
CHEN, J.-H., TARRY-ADKINS, J. L., HEPPOLETTE, C. A. A., PALMER, D. B. & OZANNE, S. E.
(2010). Early-life nutrition influences thymic growth in male mice that may be related
to the regulation of longevity. Clinical Science 118, 429–438.
CHEN, S. P., GOLEMBOSKI, K. A., PIEPENBRINK, M. & DIETERT, R. R. (2004). Developmental
immunotoxicity of lead in the rat: Influence of maternal diet. Journal of Toxicology
and Environmental Health, Part A 67, 495–511.
CLUBB, R., ROWCLIFFE, M., LEE, P., MAR, K. U., MOSS, C. & MASON, G. J. (2008).
Compromised survivorship in zoo elephants. Science 322, 1649–1649.
COHEN, S., DANZAKI, K. & MACIVER, N. J. (2017). Nutritional effects on T-cell
immunometabolism. European Journal of Immunology 47, 225–235.
COTTER, S. C., SIMPSON, S. J., RAUBENHEIMER, D. & WILSON, K. (2011). Macronutrient
balance mediates trade-offs between immune function and life history traits.
Functional Ecology 25, 186–198.
COUSSENS, L. M. & WERB, Z. (2002). Inflammation and cancer. Nature 420, 860–867.
DANDONA, P., ALJADA, A. & BANDYOPADHYAY, A. (2004). Inflammation: the link between
insulin resistance, obesity and diabetes. Trends in Immunology 25, 4–7.
DAS, U. N. (2001). Is obesity an inflammatory condition? Nutrition 17, 953–966.
DE HEREDIA, F. P., GÓMEZ-MARTÍNEZ, S. & MARCOS, A. (2012). Obesity, inflammation and
the immune system. Proceedings of the Nutrition Society 71, 332–338.
DENG, F. E., SHIVAPPA, N., TANG, Y., MANN, J. R. & HEBERT, J. R. (2016). Association
between diet-related inflammation, all-cause, all-cancer, and cardiovascular disease
31
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
mortality, with special focus on prediabetics: findings from NHANES III. European
Journal of Nutrition 56, 1085–1093.
DONZIS, E. J. & TRONSON, N. C. (2014). Modulation of learning and memory by cytokines:
signaling mechanisms and long term consequences. Neurobiology of Learning and
Memory 115, 68–77.
DUVAL, S. (2005). The trim and fill method. In Publication bias in meta-analysis (ed. H. R.
Rothstein, A. J. Sutton and M. Borenstein), pp. 127–144. John Wiley & Sons, Ltd,
Chichester.
EGGER, M., SMITH, G. D., SCHNEIDER, M. & MINDER, C. (1997). Bias in meta-analysis
detected by a simple, graphical test. British Medical Journal 315, 629–634.
ENTING, H., BOERSMA, W. J. A., CORNELISSEN, J. B. W. J., VAN WINDEN, S. C. L.,
VERSTEGEN, M. W. A. & VAN DER AAR, P. J. (2007). The effect of low-density
broiler breeder diets on performance and immune status of their offspring. Poultry
Science 86, 282–290.
EQUILS, O., SINGH, S., KARABURUN, S., LU, D., THAMOTHARAN, M. & DEVASKAR, S. U.
(2005). Intra-uterine growth restriction downregulates the hepatic toll like receptor-4
expression and function. Clinical and Developmental Immunology 12, 59–66.
ERICSSON, J. D., JANMAAT, A. F., LOWENBERGER, C. & MYERS, J. H. (2009). Is decreased
generalized immunity a cost of Bt resistance in cabbage loopers Trichoplusia ni?
Journal of Invertebrate Pathology 100, 61–67.
ESQUIFINO, A. I., CANO, P., JIMENEZ-ORTEGA, V., FERNÁNDEZ-MATEOS, M. P. &
CARDINALI, D. P. (2007). Immune response after experimental allergic
encephalomyelitis in rats subjected to calorie restriction. Journal of
Neuroinflammation 4, 6.
32
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
EVANS, M. L., WILKE, N. F., O'REILLY, P. T. & FLEMING, I. A. (2014). Transgenerational
effects of parental rearing environment influence the survivorship of captive-born
offspring in the wild. Conservation Letters 7, 371–379.
FADEL, S. & SARZOTTI, M. (2000). Cellular immune responses in neonates. International
Reviews of Immunology 19, 173–193.
FANTUZZI, G. (2005). Adipose tissue, adipokines, and inflammation. Journal of Allergy and
Clinical Immunology 115, 911–919.
FERNANDES, G. (2008). Progress in nutritional immunology. Immunologic Research 40, 244–
261.
FERNÁNDEZ-SÁNCHEZ, A., MADRIGAL-SANTILLÁN, E., BAUTISTA, M., ESQUIVEL-SOTO, J.,
MORALES-GONZÁLEZ, Á., ESQUIVEL-CHIRINO, C., DURANTE-MONTIEL, I., SÁNCHEZ-
RIVERA, G., VALADEZ-VEGA, C. & MORALES-GONZÁLEZ, J. A. (2011).
Inflammation, oxidative stress, and obesity. International Journal of Molecular
Sciences 12, 3117.
FLACKE, G. L., TKALČIĆ, S., STECK, B., WARREN, K. & MARTIN, G. B. (2016). A
retrospective analysis of mortality in captive pygmy hippopotamus (Choeropsis
liberiensis) from 1912 to 2014. Zoo Biology 35, 556–569.
GALIC, S., OAKHILL, J. S. & STEINBERG, G. R. (2010). Adipose tissue as an endocrine organ.
Molecular and Cellular Endocrinology 316, 129–139.
GARCIA, A. M., FADEL, S. A., CAO, S. & SARZOTTI, M. (2000). T cell immunity in neonates.
Immunologic Research 22, 177–190.
GEBHARDT, B. M. & NEWBERNE, P. M. (1974). Nutritional and immunological
responsiveness - T-cell function in offspring of lipotrope-deficient and protein-
deficient rats. Immunology 26, 489–495.
33
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
GIBSON, D. L., GILL, S. K., BROWN, K., TASNIM, N., GHOSH, S., INNIS, S. & JACOBSON, K.
(2015). Maternal exposure to fish oil primes offspring to harbor intestinal pathobionts
associated with altered immune cell balance. Gut Microbes 6, 24–32.
GLESER, L. J. & OLKIN, I. (2009). Stochastically dependent effect sizes. In The handbook of
research synthesis and meta-analysis (ed. H. Cooper, L. V. Hedges and J. C.
Valentine), pp. 357–376. Russell Sage Foundation, New York.
GODFREY, K. M., GLUCKMAN, P. D. & HANSON, M. A. (2010). Developmental origins of
metabolic disease: life course and intergenerational perspectives. Trends in
Endocrinology & Metabolism 21, 199–205.
GODFREY, K. M., LILLYCROP, K. A., BURDGE, G. C., GLUCKMAN, P. D. & HANSON, M. A.
(2007). Epigenetic mechanisms and the mismatch concept of the developmental
origins of health and disease. Pediatric Research 61, 5R–10R.
GOOD, R. A. & LORENZ, E. (1992). Nutrition and cellular immunity. International Journal of
Immunopharmacology 14, 361–366.
GOSBY, A. K., CONIGRAVE, A. D., RAUBENHEIMER, D. & SIMPSON, S. J. (2014). Protein
leverage and energy intake. Obesity Reviews 15, 183–191.
GRAYSON, B. E., LEVASSEUR, P. R., WILLIAMS, S. M., SMITH, M. S., MARKS, D. L. &
GROVE, K. L. (2010). Changes in melanocortin expression and inflammatory
pathways in fetal offspring of nonhuman primates fed a high-fat diet. Endocrinology
151, 1622–1632.
GRINDSTAFF, J. L., BRODIE, E. D. & KETTERSON, E. D. (2003). Immune function across
generations: integrating mechanism and evolutionary process in maternal antibody
transmission. Proceedings of the Royal Society B: Biological Sciences 270, 2309–
2319.
34
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
GRINDSTAFF, J. L., DEMAS, G. E. & KETTERSON, E. D. (2005). Diet quality affects egg size
and number but does not reduce maternal antibody transmission in Japanese quail
Coturnix japonica. Journal of Animal Ecology 74, 1051–1058.
GRISSOM, R. J. & KIM, J. J. (2001). Review of assumptions and problems in the appropriate
conceptualization of effect size. Psychological Methods 6, 135–146.
GRUEBER, C. E., REID-WAINSCOAT, E. E., FOX, S., BELOV, K., SHIER, D. M., HOGG, C. J. &
PEMBERTON, D. (2017). Increasing generations in captivity is associated with
increased vulnerability of Tasmanian devils to vehicle strike following release to the
wild. Scientific Reports 7, 2161.
GUREVITCH, J., CURTIS, P. S. & JONES, M. H. (2001). Meta-analysis in ecology. Advances in
Ecological Research 32, 199–247.
HADFIELD, J. & NAKAGAWA, S. (2010). General quantitative genetic methods for
comparative biology: phylogenies, taxonomies and multi-trait models for continuous
and categorical characters. Journal of Evolutionary Biology 23, 494–508.
HAN, T. S., SATTAR, N. & LEAN, M. (2006). Assessment of obesity and its clinical
implications. British Medical Journal 333, 695–698.
HE, Z. X., SUN, Z. H., YANG, W. Z., BEAUCHEMIN, K. A., TANG, S. X., ZHOU, C. S., HAN, X.
F., WANG, M., KANG, J. H. & TAN, Z. L. (2014). Effects of maternal protein or
energy restriction during late gestation on immune status and responses to
lipopolysaccharide challenge in postnatal young goats. Journal of Animal Science 92,
4856–4864.
HEDGES, L. (1981). Distributional theory for Glass's estimator of effect size and related
estimators. Journal of Educational Statistics 6, 107–128.
35
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
HEILBRONN, L. K. & CAMPBELL, L. V. (2008). Adipose tissue macrophages, low grade
inflammation and insulin resistance in human obesity. Current Pharmaceutical
Design 14, 1225–1230.
HERMAN, J. J. & SULTAN, S. E. (2016). DNA methylation mediates genetic variation for
adaptive transgenerational plasticity. Proceedings of the Royal Society B: Biological
Sciences 283, 20160988.
HIBSHMAN, J. D., HUNG, A. & BAUGH, L. R. (2016). Maternal diet and insulin-like signaling
control intergenerational plasticity of progeny size and starvation resistance. PLoS
Genetics 12, e1006396.
HIGGINS, J. P. T. & THOMPSON, S. G. (2002). Quantifying heterogeneity in a meta-analysis.
Statistics in Medicine 21, 1539–1558.
HIGGINS, J. P. T., THOMPSON, S. G., DEEKS, J. J. & ALTMAN, D. G. (2003). Measuring
inconsistency in meta-analyses. British Medical Journal 327, 557–560.
HO, S.-M. (2010). Environmental epigenetics of asthma – an update. The Journal of Allergy
and Clinical Immunology 126, 453–465.
HONG, L. A., ZHANG, J. J. & ZHANG, Z. B. (2004). Food restriction in pregnant rat-like
hamsters (Cricetulus triton) affects endocrine, immune function and odor
attractiveness of male offspring. Physiology & Behavior 82, 453–458.
HUNT, J., BROOKS, R., JENNIONS, M. D., SMITH, M. J., BENTSEN, C. L. & BUSSIERE, L. F.
(2004). High-quality male field crickets invest heavily in sexual display but die
young. Nature 432, 1024–1027.
IBRAHIM, A. M. A. & KIM, Y. (2006). Parasitism by Cotesia plutellae alters the hemocyte
population and immunological function of the diamondback moth, Plutella xylostella.
Journal of Insect Physiology 52, 943–950.
36
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
IWASA, T., MATSUZAKI, T., TUNGALAGSUVD, A., MUNKHZAYA, M., KUWAHARA, A., YASUI,
T. & IRAHARA, M. (2015). Prenatal undernutrition increases the febrile response to
lipopolysaccharides in adulthood in male rats. International Journal of
Developmental Neuroscience 44, 1–5.
JENNIONS, M. D., LORTIE, C. J., ROSENBERG, M. S. & ROTHSTEIN, H. R. (2013). Publication
and related biases. In Handbook of meta-analysis in ecology and evolution. (ed. J.
Koricheva, J. Gurevitch and K. Mengersen), pp. 207–236. Princeton University Press,
Princeton NJ.
KANG, S. S., KURTI, A., FAIR, D. A. & FRYER, J. D. (2014). Dietary intervention rescues
maternal obesity induced behavior deficits and neuroinflammation in offspring.
Journal of Neuroinflammation 11, 156.
KANGASSALO, K., VALTONEN, T. M., ROFF, D., POLKKI, M., DUBOVSKIY, I. M., SORVARI, J.
& RANTALA, M. J. (2015). Intra- and trans-generational effects of larval diet on
susceptibility to an entomopathogenic fungus, Beauveria bassiana, in the greater wax
moth, Galleria mellonella. Journal of Evolutionary Biology 28, 1453–1464.
KEIGHTLEY, M. C., WONG, R. B. M. & LIESCHKE, G. J. (2013). Immune priming: mothering
males modulate immunity. Current Biology 23, R76–R78.
KLENGEL, T., DIAS, B. G. & RESSLER, K. J. (2016). Models of intergenerational and
transgenerational transmission of risk for psychopathology in mice.
Neuropsychopharmacology 41, 219–231.
KNORR, E., SCHMIDTBERG, H., ARSLAN, D., BINGSOHN, L. & VILCINSKAS, A. (2015).
Translocation of bacteria from the gut to the eggs triggers maternal transgenerational
immune priming in Tribolium castaneum. Biology Letters 11, 20150885.
37
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
KOZAK, W., KLUGER, M. J., SOSZYNSKI, D., CONN, C. A., RUDOLPH, K., LEON, L. R. &
ZHENG, H. U. I. (1998). IL-6 and IL-1β in fever: studies using cytokine-deficient
(knockout) mice. Annals of the New York Academy of Sciences 856, 33–47.
LANDGRAF, M. A., LANDGRAF, R. G., SILVA, R. C., SEMEDO, P., CÂMARA, N. O. S. &
FORTES, Z. B. (2012). Intrauterine undernourishment alters TH1/TH2 cytokine
balance and attenuates lung allergic inflammation in wistar rats. Cellular Physiology
and Biochemistry 30, 552–562.
LANDGRAF, M. A., TOSTES, R. D. C. A., BORELLI, P., TENORIO ZORN, T. M., NIGRO, D.,
CARVALHO, M. H. C. & FORTES, Z. B. (2007). Mechanisms involved in the reduced
leukocyte migration in intrauterine undernourishment. Nutrition 23, 145–156.
LANGLEY, S. C., SEAKINS, M., GRIMBLE, R. F. & JACKSON, A. A. (1994). The acute phase
response of adult rats is altered by in utero exposure to maternal low protein diets.
Journal of Nutrition 124, 1588–1596.
LAVALLARD, V. J., MEIJER, A. J., CODOGNO, P. & GUAL, P. (2012). Autophagy, signaling
and obesity. Pharmacological Research 66, 513–525.
LAZZARO, B. P. & LITTLE, T. J. (2009). Immunity in a variable world. Philosophical
Transactions of the Royal Society B: Biological Sciences 364, 15–26.
LE COUTEUR, D. G., SOLON-BIET, S., COGGER, V. C., MITCHELL, S. J., SENIOR, A., DE
CABO, R., RAUBENHEIMER, D. & SIMPSON, S. J. (2016). The impact of low-protein
high-carbohydrate diets on aging and lifespan. Cellular and Molecular Life Sciences
73, 1237–1252.
LEE, K. P., SIMPSON, S. J., CLISSOLD, F. J., BROOKS, R., BALLARD, J. W. O., TAYLOR, P. W.,
SORAN, N. & RAUBENHEIMER, D. (2008). Lifespan and reproduction in Drosophila:
new insights from nutritional geometry. Proceedings of the National Academy of
Sciences 105, 2498–2503.
38
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
LOCHMILLER, R. L. & DEERENBERG, C. (2000). Trade-offs in evolutionary immunology: just
what is the cost of immunity? Oikos 88, 87–98.
LORENZ, L. M. & KOELLA, J. C. (2011). Maternal environment shapes the life history and
susceptibility to malaria of Anopheles gambiae mosquitoes. Malaria Journal 10, 382.
MA, X., TORBENSON, M., HAMAD, A. R. A., SOLOSKI, M. J. & LI, Z. (2008). High-fat diet
modulates non-CD1d-restricted natural killer T cells and regulatory T cells in mouse
colon and exacerbates experimental colitis. Clinical & Experimental Immunology
151, 130–138.
MACHOVSKY-CAPUSKA, G. E., SENIOR, A. M., SIMPSON, S. J. & RAUBENHEIMER, D. (2016).
The multidimensional nutritional niche. Trends in Ecology & Evolution 31, 355–365.
MAGRONE, T. & JIRILLO, E. (2015). Childhood obesity: immune response and nutritional
approaches. Frontiers in Immunology 6, 76.
MAJDE, J. A. & KRUEGER, J. M. (2005). Links between the innate immune system and sleep.
Journal of Allergy and Clinical Immunology 116, 1188–1198.
MANZEL, A., MULLER, D. N., HAFLER, D. A., ERDMAN, S. E., LINKER, R. A. &
KLEINEWIETFELD, M. (2014). Role of “Western Diet” in inflammatory autoimmune
diseases. Current allergy and asthma reports 14, 404–404.
MØLLER, A. P. & JENNIONS, M. D. (2001). Testing and adjusting for publication bias. Trends
in Ecology & Evolution 16, 580–586.
MUÑOZ-MARTINEZ, E., MARCOS, A., VARELA, P., UNZAGA, M. T. & VARELA, G. (1986).
Additive effects of protein-malnutrition and cortisol on plasma-proteins and
immunoglobulins of rat dams, and their offspring. Archivos Latinoamericanos De
Nutricion 36, 714–724.
39
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
NAKAGAWA, S. (2015). Missing data: mechanisms, methods and messages. In Ecological
statistics: contemporary theory and application (ed. G. A. Fox, S. Negrete-
Yankelevich and V. J. Sosa), pp. 81–105. Oxford University Press, Oxford.
NAKAGAWA, S. & CUTHILL, I. C. (2007). Effect size, confidence interval and statistical
significance: a practical guide for biologists. BIological Reviews 82, 591–605.
NAKAGAWA, S., NOBLE, D. W. A., SENIOR, A. M. & LAGISZ, M. (2017). Meta-evaluation of
meta-analysis: ten appraisal questions for biologists. BMC Biology 15, 18.
NAKAGAWA, S. & POULIN, R. (2012). Meta-analytic insights into evolutionary ecology: an
introduction and synthesis. Evolutionary Ecology 26, 1085–1099.
NAKAGAWA, S., POULIN, R., MENGERSEN, K., REINHOLD, K., ENGQVIST, L., LAGISZ, M. &
SENIOR, A. M. (2015). Meta-analysis of variation: ecological and evolutionary
applications and beyond. Methods in Ecology and Evolution 6, 143–152.
NAKAGAWA, S. & SANTOS, E. S. A. (2012). Methodological issues and advances in
biological meta-analysis. Evolutionary Ecology 26, 1253–1274.
NOBLE, D. W. A., LAGISZ, M., O'DEA, R. E. & NAKAGAWA, S. (2017). Nonindependence
and sensitivity analyses in ecological and evolutionary meta-analyses. Molecular
Ecology 26, 2410–2425.
PAYOLLA, T. B., LEMES, S. F., DE FANTE, T., REGINATO, A., MENDES DA SILVA, C., DE
OLIVEIRA MICHELETTI, T., RODRIGUES, H. G., TORSONI, A. S., MILANSKI, M. &
TORSONI, M. A. (2016). High-fat diet during pregnancy and lactation impairs the
cholinergic anti-inflammatory pathway in the liver and white adipose tissue of mouse
offspring. Molecular and Cellular Endocrinology 422, 192–202.
PIGEAULT, R., GARNIER, R., RIVERO, A. & GANDON, S. (2016). Evolution of
transgenerational immunity in invertebrates. Proceedings of the Royal Society B:
Biological Sciences 283, 20161136.
40
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
PÖLKKI, M., KANGASSALO, K. & RANTALA, M. J. (2012). Transgenerational effects of heavy
metal pollution on immune defense of the blow fly Protophormia terraenovae. PLoS
ONE 7, e38832.
PONTON, F., WILSON, K., COTTER, S., RAUBENHEIMER, D. & SIMPSON, S. (2011). Nutritional
immunology: a multi-dimensional approach. PLOS Pathogens 7, e1002223.
PONTON, F., WILSON, K., HOLMES, A. J., COTTER, S. C., RAUBENHEIMER, D. & SIMPSON, S.
J. (2013). Integrating nutrition and immunology: a new frontier. Journal of Insect
Physiology 59, 130–137.
POON, K., ABRAMOVA, D., HO, H. T. & LEIBOWITZ, S. (2016). Prenatal fat-rich diet exposure
alters responses of embryonic neurons to the chemokine, CCL2, in the hypothalamus.
Neuroscience 324, 407–419.
POULIN, R. & THOMAS, F. (2008). Epigenetic effects of infection on the phenotype of host
offspring: parasites reaching across host generations. Oikos 117, 331–335.
POVEY, S., COTTER, S. C., SIMPSON, S. J. & WILSON, K. (2014). Dynamics of macronutrient
self-medication and illness-induced anorexia in virally infected insects. Journal of
Animal Ecology 83, 245–255.
R CORE TEAM (2016). R: A language and environment for statistical computing, Vienna.
RAUBENHEIMER, D. & SIMPSON, S. J. (1993). The geometry of compensatory feeding in the
locust. Animal Behaviour 45, 953–964.
RAUBENHEIMER, D., SIMPSON, S. J., LE COUTEUR, D. G., SOLON-BIET, S. M. & COOGAN, S.
C. P. (2016). Nutritional ecology and the evolution of aging. Experimental
Gerontology 86, 50–61.
RAWSKI, M. & JÓZEFIAK, D. (2014). Body condition scoring and obesity in captive African
side-neck turtles (Pelomedusidae). Annals of Animal Science 14, 573–584.
41
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
REYNOLDS, C. M., LI, M., GRAY, C. & VICKERS, M. H. (2014). Early-life growth hormone
treatment to offspring of undernourished mothers alters metabolic parameters in
primary adipocytes in adulthood. Growth Factors 32, 34–40.
ROBERTSON, L. T. & MITCHELL, J. R. (2013). Benefits of short-term dietary restriction in
mammals. Experimental Gerontology 48, 1043–1048.
ROSE, P. E. & ROFFE, S. M. (2013). A case study of Malayan tapir (Tapirus indicus)
husbandry practice across 10 zoological collections. Zoo Biology 32, 347–356.
ROSENBERG, M. S., ROTHSTEIN, H. R. & GUREVITCH, J. (2013). Effect sizes: conventional
choices and calculations. In Handbook of meta-analysis in ecology and evolution (ed.
J. Koricheva, J. Gurevitch and K. Mengersen), pp. 284–299. Princeton University
Press, Princeton NJ.
ROTHSTEIN, H. R., SUTTON, A. J. & BORENSTEIN, M. (2005). Publication bias in meta-
analysis: prevention, assessment and adjustments. Wiley, Chichester.
SEGOVIA, S. A., VICKERS, M. H., ZHANG, X. D., GRAY, C. & REYNOLDS, C. M. (2015).
Maternal supplementation with conjugated linoleic acid in the setting of diet-induced
obesity normalises the inflammatory phenotype in mothers and reverses metabolic
dysfunction and impaired insulin sensitivity in offspring. Journal of Nutritional
Biochemistry 26, 1448–1457.
SENIOR, A. M., GRUEBER, C. E., KAMIYA, T., LAGISZ, M., O'DWYER, K., SANTOS, E. S. A. &
NAKAGAWA, S. (2016a). Heterogeneity in ecological and evolutionary meta-analyses:
its magnitude and implications. Ecology 97, 3293–3299.
SENIOR, A. M., GRUEBER, C. E., MACHOVSKY-CAPUSKA, G., SIMPSON, S. J. &
RAUBENHEIMER, D. (2016b). Macronutritional consequences of food generalism in an
invasive mammal, the wild boar. Mammalian Biology - Zeitschrift für Säugetierkunde
81, 523–526.
42
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
SENIOR, A. M., NAKAGAWA, S., LIHOREAU, M., SIMPSON, S. J. & RAUBENHEIMER, D.
(2015). An overlooked consequence of dietary mixing: a varied diet reduces
interindividual variance in fitness. The American Naturalist 186, 649–659.
SHAMSEER, L., MOHER, D., CLARKE, M., GHERSI, D., LIBERATI, A., PETTICREW, M.,
SHEKELLE, P. & STEWART, L. A. (2015). Preferred reporting items for systematic
review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation.
British Medical Journal 349, g7647.
SHANLEY, D. P., AW, D., MANLEY, N. R. & PALMER, D. B. (2009). An evolutionary
perspective on the mechanisms of immunosenescence. Trends in Immunology 30,
374–381.
SHARKEY, D., SYMONDS, M. E. & BUDGE, H. (2009). Adipose tissue inflammation:
developmental ontogeny and consequences of gestational nutrient restriction in
offspring. Endocrinology 150, 3913–3920.
SHIKANO, I., OAK, M. C., HALPERT-SCANDERBEG, O. & CORY, J. S. (2015). Trade-offs
between transgenerational transfer of nutritional stress tolerance and immune priming.
Functional Ecology 29, 1156–1164.
SIMPSON, S. J., CLISSOLD, F. J., LIHOREAU, M., PONTON, F., WILDER, S. M. &
RAUBENHEIMER, D. (2015). Recent advances in the integrative nutrition of
arthropods. Annual Review of Entomology 60, 293–311.
SIMPSON, S. J. & RAUBENHEIMER, D. (2005). Obesity: the protein leverage hypothesis.
Obesity Reviews 6, 133–142.
SIMPSON, S. J. & RAUBENHEIMER, D. (2012). The nature of nutrition: a unifying framework
from animal adaptation to human obesity. Princeton University Press.
SKINNER, M. K. (2008). What is an epigenetic transgenerational phenotype?: F3 or F2.
Reproductive Toxicology 25, 2–6.
43
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
SOLON-BIET, S. M., MCMAHON, A. C., BALLARD, J. W. O., RUOHONEN, K., WU, L. E.,
COGGER, V. C., WARREN, A., HUANG, X., PICHAUD, N., MELVIN, R. G., GOKARN, R.,
KHALIL, M., TURNER, N., COONEY, G. J., SINCLAIR, D. A., RAUBENHEIMER, D.,
LE COUTEUR, D. G. & SIMPSON, S. J. (2014). The ratio of macronutrients, not caloric
intake, dictates cardiometabolic health, aging, and longevity in ad libitum-fed mice.
Cell Metabolism 19, 418–430.
STERNBERG, E. D., DE ROODE, J. C. & HUNTER, M. D. (2015). Trans-generational parasite
protection associated with paternal diet. Journal of Animal Ecology 84, 310–321.
STEWART, T., JUNG, F. F., MANNING, J. & VEHASKARI, V. M. (2005). Kidney immune cell
infiltration and oxidative stress contribute to prenatally programmed hypertension.
Kidney International 68, 2180–2188.
TARRY-ADKINS, J. L., CHEN, J. H., JONES, R. H., SMITH, N. H. & OZANNE, S. E. (2010).
Poor maternal nutrition leads to alterations in oxidative stress, antioxidant defense
capacity, and markers of fibrosis in rat islets: Potential underlying mechanisms for
development of the diabetic phenotype in later life. Faseb Journal 24, 2762–2771.
TRIGGS, A. M. & KNELL, R. J. (2012). Parental diet has strong transgenerational effects on
offspring immunity. Functional Ecology 26, 1409–1417.
TUCHSCHERER, M., OTTEN, W., KANITZ, E., GRABNER, M., TUCHSCHERER, A., BELLMANN,
O., REHFELDT, C. & METGES, C. C. (2012). Effects of inadequate maternal dietary
protein:carbohydrate ratios during pregnancy on offspring immunity in pigs. BMC
Veterinary Research 8, 232.
VALTONEN, T. M., KANGASSALO, K., PÖLKKI, M. & RANTALA, M. J. (2012).
Transgenerational effects of parental larval diet on offspring development time, adult
body size and pathogen resistance in Drosophila melanogaster. PLoS ONE 7, e31611.
44
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
VAN BUUREN, S. & GROOTHUIS-OUDSHOORN, K. (2011). mice: multivariate imputation by
chained equations in R. Journal of Statistical Software 45, 1–67.
VAN DER KOOY, K., LEENEN, R., SEIDELL, J. C., DEURENBERG, P. & HAUTVAST, J. G.
(1993). Effect of a weight cycle on visceral fat accumulation. The American Journal
of Clinical Nutrition 58, 853–857.
VIECHTBAUER, W. (2010). Conducting meta-analyses in R with the metafor package. Journal
of Statistical Software 36, 1–48.
WANG, H., XUE, Y., WANG, B., ZHAO, J., YAN, X., HUANG, Y., DU, M. & ZHU, M. J. (2014).
Maternal obesity exacerbates insulitis and type 1 diabetes in non-obese diabetic mice.
Reproduction 148, 73–79.
WEIGLE, D. S., BREEN, P. A., MATTHYS, C. C., CALLAHAN, H. S., MEEUWS, K. E., BURDEN,
V. R. & PURNELL, J. Q. (2005). A high-protein diet induces sustained reductions in
appetite, ad libitum caloric intake, and body weight despite compensatory changes in
diurnal plasma leptin and ghrelin concentrations. The American Journal of Clinical
Nutrition 82, 41–48.
WHITE, C. L., PISTELL, P. J., PURPERA, M. N., GUPTA, S., FERNANDEZ-KIM, S. O., HISE, T.
L., KELLER, J. N., INGRAM, D. K., MORRISON, C. D. & BRUCE-KELLER, A. J. (2009).
Effects of high fat diet on Morris maze performance, oxidative stress, and
inflammation in rats: Contributions of maternal diet. Neurobiology of Disease 35, 3–
13.
WILLIAMS, S. E. & HOFFMAN, E. A. (2009). Minimizing genetic adaptation in captive
breeding programs: a review. Biological Conservation 142, 2388–2400.
XUE, Y., WANG, H., DU, M. & ZHU, M.-J. (2014). Maternal obesity induces gut
inflammation and impairs gut epithelial barrier function in nonobese diabetic mice.
Journal of Nutritional Biochemistry 25, 758–764.
45
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
YAN, X., HUANG, Y., WANG, H., DU, M., HESS, B. W., FORD, S. P., NATHANIELSZ, P. W. &
ZHU, M. J. (2011). Maternal obesity induces sustained inflammation in both fetal and
offspring large intestine of sheep. Inflammatory Bowel Diseases 17, 1513–1522.
YOKOMIZO, H., INOGUCHI, T., SONODA, N., SAKAKI, Y. & TAKAYANAGI, R. (2014). Maternal
high fat diet induced insulin resistance and deterioration of pancreatic beta cell
function with a gender difference in mature offspring. Diabetologia 56, S271–S271.
ZIMMET, P., ALBERTI, K. G. M. M. & SHAW, J. (2001). Global and societal implications of
the diabetes epidemic. Nature 414, 782–787.
VIII. SUPPORTING INFORMATION
Additional supporting information may be found in the online version of this article.
Appendix S1. Data extracted from the 38 studies used in the meta-analysis.
Appendix S2. R code used for conducting analyses.
Appendix S3. Supplementary methods, figures and tables
(1) Multiple imputation
(2) Non-independence
(3) Tests of publication bias
(4) Supplementary figures
(5) Supplementary tables
46
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
FIGURES
Fig. 1. PRISMA flow diagram of data collection and selection process (N indicates the
number of papers, except in ‘Study type’ where N indicates the number of effect sizes
estimated, including those with missing standard deviations).
47
1129
1130
1131
1132
1133
Fig. 2. Nutritional geometry (NG) is a multidimensional framework for conceptualising
foods, nutritional state and nutrient requirements (see Simpson & Raubenheimer, 2012 for a
review). Using NG, n-nutrients can be displayed in an n-dimensional Cartesian space, where
each coordinate gives an amount and balance of the nutrients modelled. For instance, here we
consider kilojoules of protein and non-protein energy on the x- and y-axes, respectively. An
individual’s state is modelled as their position in this ‘nutrient space’, and moves as foods are
eaten. Using NG, foods are represented by ‘food rails’ (here, solid lines) which have a slope
denoting the nutritional composition of the food with respect to the nutrients modelled. An
individual’s state can move along or in parallel to these food rails as it eats. The requirements
of the organism are given by an ‘intake-target’ (here, a black circle). The intake target is a
state within the nutrient space which an organism would attempt to reach under free choice.
48
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
Data suggest that organisms attempt to reach a state that maximises evolutionary fitness.
Here, the food rail given in black would allow the organism to reach the intake target; an
optimal food. The upper and lower grey lines contain less and more protein relative to non-
protein than the optimal food, respectively (grey arrows). The position of an individual’s state
along a food rail may be downwardly manipulated by restricting feeding or bulking out food
with non-digestible components, or upwardly biased by increasing the energy density of a
food (black arrows).
49
1146
1147
1148
1149
1150
1151
1152
Fig. 3. Intergenerational effects of parental diet on eight measures of immunological function. On each panel, effects sizes (mean, points, and
95% confidence interval, lines) are quantified as d, the mean difference in offspring immune response from parents on treatment diets relative to
offspring of parents on control diets (increased d indicates increased immunological measures). N values indicate the number of effect sizes; K
50
1153
1154
1155
values indicate the number of studies. Overall means are produced by a multilevel meta-analytic model that accounts for between-study
variance.
51
1156
1157
Table 1. List of 38 studies from which data were extracted.
Reference Study species (order) N1
Bilbo & Tsang (2010) Rattus norvegicus (Rodentia) 8
Blackmore et al. (2012) Mus musculus (Rodentia) 1
Boots & Roberts (2012) Plodia interpunctella (Lepidoptera) 8
Calder & Yaqoob (2000) Rattus norvegicus (Rodentia) 30
Chadio et al. (2016) Ovis aries (Artiodactyla) 6
Chen et al. (2004) Rattus norvegicus (Rodentia) 8
Chen et al. (2009) Mus musculus (Rodentia) 2
Chen et al. (2010) Mus musculus (Rodentia) 11
Enting et al. (2007) Gallus gallus (Galliformes) 48
Equils et al. (2005) Rattus norvegicus (Rodentia) 2
Gebhardt & Newberne (1974) Rattus norvegicus (Rodentia) 3
Gibson et al. (2015) Rattus norvegicus (Rodentia) 7
Grayson et al. (2010) Macaca fuscata (Primate) 1
Grindstaff et al. (2005) Coturnix japonica (Galliformes) 4
He et al. (2014) Capra aegagrus hircus (Artiodactyla) 76
Hong et al. (2004) Tscherskia triton2 (Rodentia) 3
Iwasa et al. (2015) Rattus norvegicus (Rodentia) 8
Kang et al. (2014) Mus musculus (Rodentia) 12
Kangassalo et al. (2015) Galleria mellonella (Lepidoptera) 1
Landgraf et al. (2007) Rattus norvegicus (Rodentia) 10
Landgraf et al. (2012) Rattus norvegicus (Rodentia) 22
Langley et al. (1994) Rattus norvegicus (Rodentia) 6
Lorenz & Koella (2011) Anopheles gambiae (Diptera) 2
Muñoz-Martinez et al. (1986) Rattus norvegicus (Rodentia) 1
Payolla et al. (2016) Mus musculus (Rodentia) 15
52
1158
Poon et al. (2016) Rattus norvegicus (Rodentia) 3
Reynolds et al. (2014) Rattus norvegicus (Rodentia) 24
Segovia et al. (2015) Rattus norvegicus (Rodentia) 4
Sharkey et al. (2009) Ovies aries (Artiodactyla) 16
Shikano et al. (2015) Trichoplusia ni (Lepidoptera) 7
Stewart et al. (2005) Rattus norvegicus (Rodentia) 4
Tarry-Adkins et al. (2010) Rattus norvegicus (Rodentia) 2
Tuchscherer et al. (2012) Sus scrofa (Artiodactyla) 48
Wang et al. (2014) Mus musculus (Rodentia) 2
White et al. (2009) Rattus norvegicus (Rodentia) 2
Xue et al. (2014) Mus musculus (Rodentia) 12
Yan et al. (2011) Ovis aries (Artiodactyla) 15
Yokomizo et al. (2014) Mus musculus (Rodentia) 10
1Number of effect sizes extracted.
2Species name as reported in Open Tree of Life (2 September 2016); referred to in cited source as, and
synonymous with, Cricetulus triton.
53
1159
1160
1161
Table 2. Number of effect sizes included in our main analysis (and percentage out of N = 226 total effects), as characterised by type of dietary
manipulation.
Energy Total
Decreased Same Increased
Protein1 Decreased N = 41 (18.1%)
Protein restriction N = 10
Caloric restriction N = 31
N = 61 (27.0%)
Protein restriction N = 60
High fat N = 1
N = 9 (4.0%)
Protein restriction N = 3
High fat N = 6 N = 111 (49.1%)
Same N = 50 (22.1%)
Caloric restriction N = 42
Carbohydrate restriction N = 8
N = 0 (0%) N = 29 (12.8%)
All high fat
N = 79 (35.0%)
Increased N = 20 (8.8%)
Caloric restriction N = 20
N = 10 (4.4%)
All high protein
N = 6 (2.7%)
High calorie N = 4
High fat N = 2 N = 36 (15.9%)
Total N = 111 (49.1%) N = 71 (31.4%) N = 44 (19.5%) N = 226 (100%)
1% calories from protein.
54
1162
1163
1164
Table 3. Heterogeneity statistics1.
Trait N2 K3 I2total I2
studies I2residual
Adaptive, B-cell associated 37 7 29.5% 29.5% 0.0%
Adaptive, T-cell associated 28 9 98.7% 95.7% 3.1%
Innate cells 16 9 78.9% 78.9% 0.0%
Pro-inflammatory cytokines 55 19 68.0% 66.1% 1.9%
Inflammation associated, other 32 12 66.2% 40.0% 26.2%
Anti-inflammatory/immune regulation 15 6 52.3% 51.5% 0.8%
Organ mass 15 7 70.3% 0.0% 70.3%
Resistance to infection 23 6 75.6% 5.3% 70.2%
1See Table S2 for MLMA estimated variance components.
2Number of effect sizes.
3Number of studies.
55
1165
1166
1167
1168