Type 1 diabetes is an autoimmune disease in which the immune system erroneously attacks and destroys the insulin-producing pancreatic beta cells (1). Circulating immunological markers and fatty acids have been linked to the development of type 1 diabetes (2–11). Immunological markers include cytokines, chemokines, and growth factors, which are soluble messenger molecules produced by various cell types. They are crucial regulators of immune functions inducing dynamic changes in the phenotype, functioning, and migration characteristics of immune cells. While being elementary for the regulation of immune cells, these factors also regulate various functions of the tissues under inflammatory signals (12, 13). Fatty acids also play a role in several physiological reactions, and their circulating levels are affected by diet and endogenous synthesis and metabolism, which again can be affected by genes, body weight, and health status, for example (14).

Alterations in immune function, such as overactivation of Th1 and Th17 immunity as well as abnormal cytokine production, have been related to the etiology of type 1 diabetes, although the detailed pathogenesis is poorly known (15, 16). Increased levels of IFN-γ, IL-17, IL-9, IL-6, IL-1β, CCL2 (MCP1), CCL4, and CXCL10 (IP-10) (3–6, 15, 17), as well as activation and phenotypic plasticity of Th17 cells have been associated with the development of islet autoimmunity and progression from the autoantibody-positive state towards clinical type 1 diabetes (18, 19). The role of Th17 plasticity in the development of type 1 diabetes was further implicated by the observation that TGF-β induced secretion of IL-9 in Th17 cells in children with recent-onset type 1 diabetes (7, 20, 21). Children with newly diagnosed type 1 diabetes had elevated levels of granulocyte macrophage-colony stimulating factor (GM-CSF), IL-2, IL-7, IL-8, IL-1β, IL-17F, IL-21, IL-23, IL-10, and IL-27 compared to control children (6, 8). Also, in vitro studies have indicated that certain cytokines are linked to type 1 diabetes. IL-17 alone or in combination with IL-1β and IFN-γ has been demonstrated to be detrimental to human islets and beta cells (15, 19, 22). Some evidence indicates that cytokine production is altered already before the appearance of islet autoimmunity: type 1 IFN-mediated innate immune system has been shown to be activated (23) and IL-32 upregulated (24). Furthermore, high maternal levels of circulating inflammatory cytokines (IFN-γ, IL-1β, CCL2, and CCL4) during pregnancy have been associated with increased risk of type 1 diabetes in the offspring (4, 17). In addition to immunological alterations, also early metabolic dysregulation of lipids and amino acids in cord blood or in infancy have been linked to type 1 diabetes (25–27). Furthermore, several abnormalities including alterations in lipid metabolism and inflammatory events were seen in children already 1 year before seroconversion (11).

Evidence from prospective cohort studies suggests that fatty acids contribute to the risk of islet autoimmunity, although the findings are partly inconsistent (9, 10, 28–31). High levels of long-chain n-3 polyunsaturated fatty acids (PUFAs) in serum or in the erythrocyte membrane might protect from islet autoimmunity (9, 10, 31), especially at an early age (10, 31). On the other hand, in the TRIGR study, we observed that high serum proportions of saturated fatty acids (SFAs) (15:0, i17:0, ai17:0, 18:0) and conjugated linoleic acid (CLA, 18:2n-7) were associated with increased risk of islet autoimmunity (30). Higher levels of some SFAs and MUFAs were also associated with an enhanced risk of islet autoimmunity in two other large-scale birth cohort studies (DIPP and TEDDY) (29, 31).

Fatty acids have been observed to affect both innate and adaptive immunity in in vitro studies and in animal models. Fatty acids may modulate immune system, e.g., by activating or downregulating inflammatory processes such as cytokine secretion (32). In general, n-3 PUFAs are considered as anti-inflammatory, and SFAs are considered as pro-inflammatory, based on in vitro and animal studies (32). In human adults, supplementation with long-chain n-3 PUFAs decreased levels of TNF-α, IL-6, and CRP according to a meta-analysis (33). To our knowledge, there is only one intervention study observing the effects of fatty acids on the participating children’s immunological markers: supplementation with DHA during pregnancy or infancy had no effect on the production of inflammatory cytokines such as IL-1β, TNF-α, or IL-12p40 in the offspring (34). As far as we are aware, no information exists on the associations between fatty acids and immunological markers during infancy, and whether these associations are different in children who develop islet autoimmunity and those who do not.

Our aim was to study the associations between serum fatty acids and immunological markers at birth and at the age of 6 and 12 months, and to study whether these associations differed between case and control children in order to define putative mechanisms related to development of islet autoimmunity. We hypothesized that n-3 PUFAs are associated with higher levels of anti-inflammatory immunological markers during infancy, while SFAs are related to pro-inflammatory immunological markers.

The median (interquartile range) age at positivity for at least two antibodies was 4.0 (1.9–6.0) years among the 95 case children. One child tested positive for at least two antibodies already at the age of 6 months and 26 children at the age of 12 months. The most common autoantibody combinations among the case children at the time of case selection were ICA and GADA (n = 26); ICA and IAA (n = 19); ICA, IAA, and GADA (n = 15); ICA, IAA, GADA, and IA-2A (n = 15); IAA and GADA (n = 10); and ICA, GADA and IA-2A (n = 6). One child had IAA, GADA, and IA-2A; one had ICA and IA-2A; one had ICA, IAA, and IA-2A; and one child had ICA only but presented with type 1 diabetes. Characteristics of the participants are presented in Table 1. Distributions of serum fatty acid proportions are presented in Supplementary Table 1 and those of immunological markers in Supplementary Table 2 .

We observed differences in the associations between serum fatty acids and immunological markers during the first year of life in case compared to control children. In general, in children who developed islet autoimmunity, SFAs showed more associations with immunological markers, while in control children, PUFAs showed stronger associations. These observations are in line with the findings of several large-scale birth cohort studies that high levels of n-3 PUFAs were associated with decreased risk of islet autoimmunity (9, 10, 31), while SFAs tended to be associated with increased risk (29–31).

It was previously observed in the TRIGR study that several SFAs (15:0, i17:0 and ai17:0, 18:0) and CLA were associated with increased risk of islet autoimmunity (30). We observed that in case children, several SFAs and CLA were associated with many immunological markers [CXCL10 (IP-10), IL-6, IL-9, IL-17, and CM-CSF] that have been linked to the development of type 1 diabetes (3–6, 8, 39). In our study, associations between most SFAs (15:0, ai17:0, 18:0, 20:0, and 24:0) and these cytokines were inverse in cord serum and at 6 months of age, but at the age of 12 months, associations between 22:0 and 24:0 with IL-17 were direct. During the first year of life, the immune system is developing rapidly, and also the dietary intake of fatty acid changes due to increasing food variety in the child’s diet, and thus associations between circulating fatty acids and immunological markers may vary by age. Our results indicate that changes in fatty acid intake and/or metabolism occur in case children, detected as increased SFA and CLA levels, concurrently with alterations in patterns of immunological markers. Whether the observed correlations in our study between SFAs and inflammatory immunological markers indicate causality is unknown. It has been shown that SFAs have direct immunomodulatory effects. They have been shown to stimulate proliferation of T cells and their differentiation to proinflammatory phenotypes Th1 and Th17 (32), which is in line with our finding at 12 months of age. On the other hand, CLA has shown capacity to downregulate pro-inflammatory cytokines, such as IL-6, TNFα, IFNγ, and IL-1β in cell culture studies (40), which is in accordance with our result that CLA was inversely associated with IL-9.

In case children, the observed inverse relations at birth and at 6 months of age between SFAs and CLA linked to islet autoimmunity and several proinflammatory immunological markers linked to type 1 diabetes development may be explained by the fact that higher levels of proinflammatory immunological markers during the first months of life may be actually beneficial, while at later age, they may indicate inflammation. At birth, the infant’s immunity is in a tolerogenic state in order to tolerate maternal antigens and stress (innate immunity and Th1 responses are suppressed). After birth, dynamic co-development between the innate and adaptive immune system as well as microbes, are needed for developing regulated immune responses against antigens and pathogens. Cytokines mediate the activation of these early innate immune responses and the training of the adaptive immune system (12).

In control children, PUFAs were more strongly associated with many more immunological markers compared to case children. Cord serum DPA was directly associated with IL-13 and IL-1α in control children. IL-13 is secreted by activated Th2 cells and it has been associated with downregulation of several macrophage and monocyte-derived pro-inflammatory cytokines (41). In the TRIGR study, it was observed that cord serum DPA (n-3 PUFA) tended to be associated with decreased risk of later islet autoimmunity (30). Three other large-scale birth cohort studies (DIPP, DAISY, and TEDDY) have showed that n-3 PUFA (ALA, EPA, DPA, or DHA) intake or their increased proportions in serum or erythrocytes were associated with a reduced risk of islet autoimmunity (9, 10, 31). In animal models and clinical trials, an n-3 PUFA-rich diet has been shown to decrease the production of inflammatory cytokines and increase the secretion of anti-inflammatory cytokines, increase regulatory T cells, and reduce T-cell proliferation, activation, and differentiation into proinflammatory Th1 and Th17 cells (32). In addition to modulating immune responses and inflammatory mediators, n-3 PUFAs could also modulate the diversity and abundance of the gut microbiota, particularly the ratio of Firmicutes and Bacteroides (42). Changes in gut microbiota such as decreased relative abundance in Firmicutes and increased relative proportion of Bacteroides have been observed already at early age in children who later on develop type 1 diabetes (43). Furthermore, n-3 PUFA may affect gut microbiota also by regulating the levels of short-chain fatty acids or short-chain fatty acid salts (42). On the other hand, gut microbiota could modulate the absorption and metabolism of n-3 PUFAs and thus its intake and function (42).

Associations between n-6 PUFA and immunological markers were heterogenous. In control children, certain n-6 PUFAs (DGLA and adrenic acid) correlated inversely with several cytokines (IL-10, TGF-α, IL-1ra, G-CSF, IL-15, and IL-6) at 6 months of age. At 12 months of age, control children showed an inverse association between LA and IL-17, out of which the latter has been linked to type 1 diabetes (15, 19, 21). In turn, in case children at the age of 6 months, n-6 LA and AA were directly associated with cytokines (IL-6, IL-9, and CXCL10), which have been previously linked to pre-clinical and clinical type 1 diabetes (3–6, 39). Altogether, our findings indicate that both n-3 and n-6 PUFAs, particularly in the early months, may be associated with the pattern of circulating cytokines, and thus possibly play a role in the maturation of the immune system and longer-term immune responses.

In our study, DMA18 also correlated inversely in case children at 6 months of age with IL-9 and GM-CSF, i.e., cytokines that have been linked to type 1 diabetes (8, 21, 39). DMA18 is a plasmalogen, which is an unusual subgroup of phospholipids in membranes. Plasmalogens have been implicated to play an important role in many cell functions, such as cell signaling and protection against reactive oxygen species. The metabolism of plasmalogens is related to both n-3 and n-6 PUFAs: plasmalogens may serve as reservoir for PUFAs, and thus, plasmalogen deficiency may affect PUFA levels, and on the other hand, the presence of PUFA is needed for the biological activity of plasmalogens (44, 45). One of the plasmalogen species is choline plasmalogen (46). Interestingly, cord blood choline-containing phospholipids (47) and unsaturated phosphatidylcholines in infancy (27) were found to be low in children who developed islet autoimmunity.

There were no clear patterns for associations between MUFAs and immunological markers. In cord serum from control children, some MUFAs showed inverse correlation: vaccenic acid with TNF-β and oleic acid with GM-CSF, IFN-α2, and IL-13. In cases at 6 months of age, vaccenic acid correlated inversely with IL-9, and further at 12 months of age, oleic acid was inversely related to IFN-γ and palmitoleic and eicosenoic acids were inversely related to CXCL1. As far as we know, there are no earlier evidence of these kinds of associations.

The strengths of the present study include a unique sample set from several different countries, and the possibility to use unthawed samples for immunological marker analysis. We analyzed a relatively large number of fatty acids and immunological markers, which enables a wider view of complex immunomodulatory networks. A limitation is that our study design does not allow us to draw conclusions about causality regarding associations between fatty acids and immunological markers, and the results might reflect some other factors related to disease etiology. A limitation of the study is the unavailability of genetic factors related to the regulation of immunological markers (48) and fatty acid metabolism, such as polymorphism in genes encoding fatty acid desaturases and elongases (14). Generalizability of our results to the whole population remains open since our study population comprised high-risk children due to their HLA-conferred risk of type 1 diabetes and having a first-degree relative/s with type 1 diabetes.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the ethics committees of all participating centers. The study was conducted according to the standards of the Declaration of Helsinki. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

SN, MM, and SV were responsible for conception and design of the study. MK, JK, SV, and OV were responsible for the acquisition of data. JH, AV, TH, MK, IE, and PA were responsible for laboratory analysis. DC and RA did statistical analysis. SN, MM, SV, and LH drafted the article with contributions from JH and MK. MK and SV are the guarantors of this work. All authors contributed to the interpretation of the data and critically reviewed and approved the version to be published.

This work was supported by National Institutes of Health (grants 1DP3DK106918-01, HD040364, HD042444, and HD051997), the Eunice Kennedy Shriver National Institute of Child Health and Development (NICHD), National Institute of Diabetes and Digestive and Kidney Diseases, Canadian Institutes of Health Research, JDRF, the Commission of the European Communities (specific RTD programme Quality of Life and Management of Living Resources, contract QLK1-2002-00372 Diabetes Prevention), the European Foundation for the Study of Diabetes/JDRF/Novo Nordisk Focused Research Grant, Academy of Finland (Centre of Excellence in Molecular Systems Immunology and Physiology Research 2012-2017, Decision No. 250114), Dutch Diabetes Research Foundation, and Finnish Diabetes Research Foundation. The study sponsors were not involved in the design of the study; the collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.