A defining characteristic of acute (i.e. wasting) pre-pubescent protein and energy deficit is depressed inflammatory immune competence, although the aetiology of this aspect of malnutrition pathology is poorly understood(Reference Woodward and Chandra1). A recent proposition suggests that this immunological phenomenon represents a regulated pathophysiology, controlled by hormones and cytokines, rather than a biologically trivial disintegrative loss of immunological control(Reference Woodward and Chandra1). In this connection, murine models that closely mimic acute paediatric malnutrition exhibit elevations in blood levels of three potent anti-inflammatory mediators, namely corticosterone(Reference Monk, Makinen and Shrum2), IL-10(Reference Hillyer, Dao and Niemiec3) and transforming growth factor (TGF)-β(Reference Hillyer, Dao and Niemiec3), in the most advanced stages of wasting pathology. Moreover, in the same pre-pubescent experimental systems, blood levels of both corticosterone(Reference Monk, Makinen and Shrum2) and IL-10(Reference Monk and Woodward4) rise in the early stages of weight loss. Importantly, high blood levels of TGF-β are also reported in the advanced stages of a completely different model of acutely protein-deficient weanling guinea pigs vaccinated with Mycobacterium tuberculosis (Reference Dai and McMurray5). Blood concentrations of hormones and cytokines represent spillover from the extravascular compartment and may be regarded as reflective, although probably not representative, of concentrations at extravascular sites of action(Reference Hillyer, Dao and Niemiec3, Reference Bienvenu, Monneret and Gutowski6). High levels of glucocorticoids and IL-10, therefore, are positioned temporally to initiate depressed inflammatory capacity in the early stages of acute malnutrition, and a potent triad including the glucocorticoids, IL-10 and TGF-β is likewise positioned to serve in a sustaining role. It is unknown whether TGF-β might also contribute to the initiation of malnutrition-associated depression in inflammatory immune competence.
TGF-β exerts potently anti-inflammatory and immune suppressive effects impacting cells of both the innate and adaptive arms of immune defence(Reference Li, Wan and Sanjabi7), and this cytokine is regarded as a dominant peripheral mediator of anti-inflammatory self-tolerance(Reference Zhang, Yi and Xia8, Reference Li and Flavell9). The objective of this investigation was to determine if the blood TGF-β concentration rises during the early stages of acute pre-pubescent malnutrition in metabolically distinct murine models known to depress inflammatory immune competence(Reference Woods and Woodward10–Reference Ha, Wong and Gray12). Additionally, a confirmatory experiment was conducted assessing blood levels of TGF-β in the advanced stages of acute weanling malnutrition.
Materials and methods
Animals and facilities
Male and female C57BL/6J mice were obtained from an in-house breeding colony. Caging and housing conditions were exactly as described previously(Reference Monk, Makinen and Shrum2, Reference Ha, Wong and Gray12), and this investigation was approved by the Animal Care Committee of the University of Guelph in accordance with the guidelines of the Canadian Council on Animal Care.
Diets and feeding protocols
At 18 d of age, the animals were weaned, individually housed and given free access for 1 d to a complete purified diet that is described elsewhere(Reference Filteau and Woodward13). At 19 d of age, the mice were randomly allocated to one of three experimental groups. An age-matched control group was given ad libitum access to the aforementioned complete purified diet. An energy-restricted group consumed the complete diet in restricted daily quantities determined by calculations that relate ad libitum food intake (g food/g body weight) to chronological age in the weanling mouse(Reference Ha, Wong and Gray12). Finally, a low-protein group was given free access to a purified diet formulated to contain 0·6 % crude protein by replacement of most of the nitrogen source of the complete diet by maize starch(Reference Woods and Woodward10). These malnutrition protocols reproduce the critical features of the paediatric human pathologies of marasmus via the restricted intake protocol and incipient kwashiorkor via the low-protein protocol(Reference Monk, Makinen and Shrum2, Reference Hillyer, Dao and Niemiec3, Reference Ha, Wong and Gray12).
Experimental design
Two experiments were performed. In the first experiment, ten animals were included in each dietary group and blood samples were taken at day 3, representative of the early stages of weight loss(Reference Monk, Makinen and Shrum2). In the second experiment, intended to be confirmatory of a previous report pertaining to the advanced stages of malnutrition(Reference Hillyer, Dao and Niemiec3), sample sizes consisted of ten animals in the age-matched control group and eight animals in both malnourished groups, and blood was taken after 14 d. Equal numbers of males and females were included in all dietary groups from both experiments.
Blood sampling procedure
Orbital plexus blood samples were collected under CO2 anaesthesia as described previously(Reference Hillyer, Dao and Niemiec3). Platelet-poor plasma was collected and stored at − 80°C as described previously(Reference Hillyer, Dao and Niemiec3).
Plasma transforming growth factor-β bioassay
The assay was based on the reduction of the proliferative activity of Mv1Lu mink lung cells (ATCC CCL-64) exposed to TGF-β and was performed as described in detail elsewhere(Reference Hillyer, Dao and Niemiec3). Recombinant human TGF-β1 (BD Biosciences, San Jose, CA, USA; catalogue #559119) was used to generate standard curves (average R 2 0·98), and only linear portions of the curves were used. The intra-assay CV averaged 4·6 %, and the detection limit was 46 pg/ml. Each plasma sample was analysed in triplicate.
Carcass composition
Carcasses were stored at − 80°C to await analysis of DM and total lipid content as described previously(Reference Woods and Woodward10).
Statistical analysis
Statistical analyses were conducted using the SAS system (SAS Institute, Cary, NC, USA) for Windows (version 9.0), and a pre-determined upper limit of probability of P ≤ 0·05 was applied for statistical significance. Data were subjected to a one-way ANOVA followed, if justified by Tukey's Studentised Range test. Datasets not exhibiting a normal distribution were transformed. Where transformation attempts were unsuccessful, data were subjected to the Kruskal–Wallis test (χ2 approximation) followed, if justified, by Wilcoxon two-sample testing.
Results
Distinct weight loss pathologies were elicited by the malnutrition protocols
Growth indices for the two experiments are shown in Table 1. Within each experiment, initial body weights did not differ among groups. Moreover, the age-matched control groups exhibited food intakes and carcass composition outcomes comparable to those reported previously for C57BL/6J weanlings consuming the complete purified diet(Reference Monk, Makinen and Shrum2, Reference Hillyer, Dao and Niemiec3, Reference Ha, Wong and Gray12). At the end of each experiment, the body weights and food intakes of the malnourished groups were lower than their corresponding age-matched control group. Also, each malnourished group exhibited a lower carcass fat level than its corresponding age-matched control, and the restricted intake protocol induced a greater decrement in carcass fat, and hence in carcass energy, than the low-protein protocol. Overall, the loss of both fat and lean tissue produced by the two malnutrition protocols was comparable to the outcomes reported previously when inflammatory immune competence was shown to be depressed in the same experimental systems(Reference Woods and Woodward10–Reference Ha, Wong and Gray12).
(Mean values with their standard errors)
C, group that consumed ad libitum the complete diet; LP, group that consumed ad libitum an isoenergetic low-protein diet; R, group that consumed the complete diet in restricted daily quantities.
a,b,c Mean values within a row with unlike superscript letters were significantly different (P ≤ 0·05) according to Tukey's Studentised Range test, unless a different statistical procedure is indicated.
* From ANOVA of squared-transformed data. Mean values are square roots of squared means.
† Kruskal–Wallis test of Wilcoxon rank sums that were as follows: C, 187; LP, 188; R, 90.
‡ From ANOVA of inverse-transformed data. Mean values are the inverse of the transformed means.
§ From ANOVA of natural log-transformed data. Mean values are antilogs of log means.
∥ Kruskal–Wallis test of Wilcoxon rank sums that were as follows: C, 215; LP, 98·5; R, 37·5.
¶ Kruskal–Wallis test of Wilcoxon rank sums that were as follows: C, 202; LP, 83·5; R, 65·5.
Plasma transforming growth factor-β bioactivity
Plasma TGF-β bioactivities are shown in Fig. 1. Samples from all groups exhibited plasma bioactivities that exceeded the detection limit of the assay. Plasma TGF-β bioactivities found in the age-matched control groups were consistent with concentrations reported previously in plasma samples from healthy adolescent and young adult mice(Reference Hillyer, Dao and Niemiec3, Reference Tanaka, Miyazaki and Takeda14). In relation to the main objective of this investigation, plasma TGF-β bioactivities were elevated in both malnourished groups after only 3 d of weight loss, and the high blood levels of this cytokine were sustained into the advanced stages of protein and energy deficit.
Plasma transforming growth factor (TGF)-β bioactivity. Weanling C57BL/6J mice, initially 19 d old, were fed ad libitum a complete purified diet (group C, age-matched controls), or the complete diet in restricted daily quantities (group R) or were given free access to an isoenergetic low-protein diet (group LP). Equal numbers of males and females contributed to samples sizes of ten in all dietary groups (upper panel, day 3), or to sample sizes of ten (group C) and eight in each malnourished group (lower panel, day 14). Bars represent mean values and the sem is shown in each case. Within each experiment, data were analysed by one-way ANOVA. Day 3: means are antilogs of natural log-transformed means; diet P = 0·0004, pooled sem = 0·16. Day 14: diet P = 0·0001, pooled sem = 339. Bars marked with an asterisk (*) differ from their corresponding age-matched control group (P ≤ 0·05) according to Tukey's Studentised Range test.
Discussion
This investigation reveals that the plasma concentration of TGF-β rises early in response to acute pre-pubescent deficits of protein and energy, and demonstrates that the high levels of this anti-inflammatory cytokine are maintained into the advanced stages of weight loss. The findings pertaining to the later stages of malnutrition confirm previous reports(Reference Monk, Makinen and Shrum2, Reference Dai and McMurray5). This outcome was apparent in two metabolically distinct murine models that mimic the critical features of paediatric marasmus and incipient kwashiorkor(Reference Monk, Makinen and Shrum2, Reference Hillyer, Dao and Niemiec3, Reference Ha, Wong and Gray12). Thus, a potently anti-inflammatory triad of soluble mediators, namely the glucocorticoids, IL-10 and TGF-β, emerges when this work is considered together with previous reports(Reference Monk, Makinen and Shrum2–Reference Dai and McMurray5) and provides a basis for understanding both initiation and maintenance of depressed inflammatory competence in acute pre-pubescent malnutrition. More broadly, this investigation provides new experimental underpinning for the concept(Reference Woodward and Chandra1) that malnutrition-associated immune depression is part of a regulated pathophysiology, the antithesis of a biologically trivial disintegration of immunological capacities.
Attempts to achieve a biologically meaningful assessment of TGF-β concentrations in blood and other tissue fluids are complicated by the existence of three mammalian isoforms of this protein mediator, designated β1–3, (Reference Graycar, Miller and Arrick15) and by the fact that much of the cytokine circulates in latent form(Reference Taylor16). Only the β1 isoform is detectable in the blood of the healthy mouse or of mice subjected to the low-protein protocol used herein, whereas both β1 and β2 isomers are reported at high levels in the blood of mice subjected to our food intake restriction protocol(Reference Hillyer, Dao and Niemiec3). It is important, therefore, that the three mammalian isoforms exhibit indistinguishable immunological influences, at least in vitro (Reference Hillyer, Dao and Niemiec3), and equal biological activity in the bioassay used herein(Reference Graycar, Miller and Arrick15). Moreover, TGF-β immunoassays(Reference Kanzler, Baumann and Schirmacher17) and bioassays(Reference Dai and McMurray5, Reference Kanzler, Baumann and Schirmacher17) frequently include an initial activation step that releases the cytokine from its latent form, thereby inflating the measurement that is interpreted as biologically active TGF-β. However, this assay strategy was not used in the present investigation. To the extent that blood cytokine concentrations reflect extravascular levels(Reference Hillyer, Dao and Niemiec3, Reference Bienvenu, Monneret and Gutowski6), therefore, the TGF-β concentrations determined herein are directly relevant to the levels of biologically active cytokine that target cells would encounter in vivo.
Numerous possibilities merit investigation as contributors to high circulating levels of TGF-β in acute malnutrition. In this connection, findings pertaining to the small intestinal epithelium and lamina propria of acutely malnourished infants and children lend no support to the possibility of an increase in the rate of cytokine synthesis(Reference Campbell, Murch and Elia18). However, a much more extensive cellular survey is needed as virtually all cells can produce TGF-β(Reference Taylor16). Likewise, reduced turnover must be considered in parallel with findings reported for other circulating proteins of acutely malnourished children, e.g. IgG class Ig(Reference Chandra and Newberne19) and some acute-phase proteins(Reference Morlese, Forrester and Badaloo20, Reference Morlese, Forrester and Jahoor21). An intriguing possibility pertains to the extracellular matrix, which appears to serve as a reservoir for TGF-β(Reference Taylor16, Reference Annes, Munger and Rifkin22). Proteolytic degradation of the extracellular matrix liberates TGF-β from its latent complex, thereby coupling matrix turnover with the activation of TGF-β(Reference Annes, Munger and Rifkin22). Thus, by exaggerating a normal physiological phenomenon, the catabolic metabolism of acute malnutrition may support a rapid and sustained elevation in blood and tissue fluid concentrations of a potent anti-inflammatory cytokine in its active form. In this connection, protein- and energy-deprived rats exhibit decreased soluble and insoluble collagen skin content in comparison to healthy controls(Reference Anasuya and Narasing Rao23).
The glucocorticoids, IL-10 and TGF-β comprise a network of partially redundant anti-inflammatory hormonal mediators effective against diverse innate and adaptive immune defence elements and widely regarded as the backbone of peripheral tolerance(Reference Li, Wan and Sanjabi7–Reference Li and Flavell9, Reference Vignali, Collison and Workman24, Reference Karagiannidis, Akdis and Holopainen25). The present investigation completes the picture with respect to this potent network in acute pre-pubescent malnutrition. It is now apparent, from this investigation and other reports(Reference Monk, Makinen and Shrum2–Reference Monk and Woodward4), that the three component mediators rise early and are sustained in unison throughout the progression of weight loss. In turn, this points to the recent proposition of a coordinated and purposeful immune regulation in which the inflammatory form of competence is supplanted by an anti-inflammatory form that is tolerant of the self-antigens catabolically released in quantity during wasting malnutrition(Reference Woodward and Chandra1, Reference Hillyer, Dao and Niemiec3). This tolerance-centred model emphasises the adaptive benefit of a reduced risk of autoimmune disease while acknowledging the cost in terms of susceptibility to infectious disease. Other anti-inflammatory mediators, e.g. IL-35(Reference Vignali, Collison and Workman24), can be accommodated within the tolerance model, and their participation can be anticipated in the non-inflammatory immune competence of acute pre-pubescent malnutrition.
Acknowledgements
The work reported herein was supported by a Postgraduate Scholarship and a Discovery Grant awarded, respectively, to J. M. M. and B. W. by the Natural Sciences and Engineering Research Council of Canada. The study idea originated with J. M. M. The study was designed jointly by J. M. M. and B. W., and the experimental work was done by J. M. M. Both J. M. M. and B. W. contributed to the interpretation of results and the preparation of the manuscript. Technical assistance provided by Ms Rinu Pazhekattu in connection with carcass analyses is gratefully acknowledged. The authors have no conflicts of interest to declare.
