Dynamic changes in immune gene co-expression networks predict development of type 1 diabetes
Ingrid Brænne, Suna Önengüt-Gümüşcü, Ruoxi Chen, Ani Manichaikul, Stephen S. Rich, Wei‐Min Chen, Charles R. Farber, Marian Rewers, Aaron Barbour, Kimberly Bautista, Judith Baxter, Daniel Felipe-Morales, Kimberly A. Driscoll, Brigitte I. Frohnert, Marisa Stahl, Patricia Gesualdo, Michelle Hoffman, Rachel Karban, Edwin Liu, Jill Norris, Stesha Peacock, Hanan Shorrosh, Andrea K. Steck, Megan Stern, Erica Villegas, Kathleen Waugh, Jorma Toppari, Olli Simell, Annika Adamsson, Suvi Ahonen, Iris Erlund, Leena Hakola, Anne Hekkala, Henna Holappa, Heikki Hyöty, Anni Ikonen, Jorma Ilonen, Sinikka Jäminki, Sanna Jokipuu, Leena Eklund Karlsson, Jukka Kero, Miia Kähönen, Mikael Knip, Minna-Liisa Koivikko, Merja Koskinen, Mirva Koreasalo, Kalle Kurppa, Jarita Kytölä, Tiina Latva-aho, Katri Lindfors, Maria Lönnrot, Elina Mäntymäki, Markus Mattila, Maija E. Miettinen, Katja Multasuo, Teija Mykkänen, Tiina Niininen, Sari Niinistö, Mia Nyblom, Sami Oikarinen, Paula Ollikainen, Zhian Othmani, Sirpa Pohjola, Petra Rajala, Jenna Rautanen, Anne Riikonen, Eija Riski, Miia Pekkola, Minna Romo, Satu Ruohonen, Satu Simell, Maija Sjöberg, Aino Stenius, Päivi Tossavainen, Mari Vähä-Mäkilä, Sini Vainionpää, Eeva Varjonen, Riitta Veijola, Irene Viinikangas, Suvi Μ. Virtanen, Jin-Xiong She, Desmond Schatz, Diane Hopkins, Leigh Steed, Jennifer Bryant, Katherine Silvis, Michael Haller, Melissa Gardiner, Richard McIndoe, Ashok Sharma, Stephen W. Anderson, Laura M. Jacobsen, John Marks, P.d. Towe, Anette G. Ziegler, Ezio Bonifacio, Anita Gavrisan, Cigdem Gezginci, Anja Heublein, Verena HoffmannScientific2021 · DOI
Significant progress has been made in elucidating genetic risk factors influencing Type 1 diabetes (T1D); however, features other than genetic variants that initiate and/or accelerate islet autoimmunity that lead to the development of clinical T1D remain largely unknown. We hypothesized that genetic and environmental risk factors can both contribute to T1D through dynamic alterations of molecular interactions in physiologic networks. To test this hypothesis, we utilized longitudinal blood transcriptomic profiles in The Environmental Determinants of Diabetes in the Young (TEDDY) study to generate gene co-expression networks. In network modules that contain immune response genes associated with T1D, we observed highly dynamic differences in module connectivity in the 600 days (~ 2 years) preceding clinical diagnosis of T1D. Our results suggest that gene co-expression is highly plastic and that connectivity differences in T1D-associated immune system genes influence the timing and development of clinical disease.
Introduction
Type 1 Diabetes (T1D) is an autoimmune disease characterized by the destruction of insulin-producing beta cells. While genetic risk factors, particularly HLA genotypes, are well-documented, the precise molecular events leading from predisposition to clinical onset remain elusive. A significant challenge in T1D research is the "post-diagnosis" bias; most studies examine patients after the immune system has already caused irreversible damage. Furthermore, traditional transcriptomic analyses often fail to detect subtle changes in individual gene expression that precede clinical symptoms. This study addresses these gaps by shifting the focus from individual gene expression to the dynamic "wiring" of gene co-expression networks during the pre-clinical phase.
Methods
Researchers utilized data from the The Environmental Determinants of Diabetes in the Young (TEDDY) study, analyzing 1,921 whole blood transcriptomic profiles from 385 children (60 cases, 325 controls) over a 600-day window preceding diagnosis.
Overview of the analytical pipeline for identifying T1D-specific gene co-expression modules and differentially connected genes.
The study employed Weighted Gene Co-expression Network Analysis (WGCNA) to organize genes into functional modules. The primary metric used was Modular Differential Connectivity (MDC), which measures the density of interactions within a module. By comparing the connectivity ratio between T1D cases and controls across 10-day intervals, the team could identify periods of intense molecular reorganization. These modules were further prioritized based on their enrichment with known T1D risk genes identified through Genome-Wide Association Studies (GWAS).
Key Insights and Results
Connectivity as a Sensitive Amplifier
The most significant finding was that while individual gene expression differences between cases and controls were often negligible (less than 5%), the internal connectivity of immune modules increased by as much as 10-fold. This indicates that network topology is a far more sensitive biomarker for early disease progression than simple expression levels.
Maximum and minimum Modular Differential Connectivity (MDC) values across gene modules in females, highlighting significant rewiring in the purple and blue modules.
The immune system's molecular wiring is highly plastic, exhibiting "waves" of connectivity that peak and subside before diagnosis. Two modules were particularly noteworthy:
Purple Module: Linked to monocytes, macrophages, and viral response pathways (interferon signaling). In females, this module showed a massive 10-fold connectivity spike approximately 231 days before diagnosis.
Blue Module: Associated with neutrophil signatures.
Sexual Dimorphism and Hub Genes
The study observed distinct sexual dimorphism in the timing and magnitude of these network shifts. In girls, connectivity peaks occurred earlier and were generally higher in magnitude, aligning with clinical observations of faster disease progression in females. Key "hub" genes driving this rewiring included TLR8, IFIH1, and SIGLEC1.
Discussion and Limitations
By mapping GWAS results onto these dynamic networks, the authors provided functional context for genetic risk factors, demonstrating that these genes act as active hubs in the networks that fail prior to T1D onset.
However, several limitations remain. The TEDDY cohort is restricted to children with high-risk HLA genotypes, and it is unclear if these dynamics apply to the broader population or adult-onset T1D. Additionally, as half of the cases had already developed autoantibodies (seroconverted) at the start of the study window, future research is needed to investigate even earlier time points to capture the very first initiation of autoimmunity. Finally, while the link between network changes and T1D is clear, determining whether this rewiring is a primary cause or an early systemic response remains a challenge for future functional studies.
Summary updated
Abstract
Significant progress has been made in elucidating genetic risk factors influencing Type 1 diabetes (T1D); however, features other than genetic variants that initiate and/or accelerate islet autoimmunity that lead to the development of clinical T1D remain largely unknown. We hypothesized that genetic and environmental risk factors can both contribute to T1D through dynamic alterations of molecular interactions in physiologic networks. To test this hypothesis, we utilized longitudinal blood transcriptomic profiles in The Environmental Determinants of Diabetes in the Young (TEDDY) study to generate gene co-expression networks. In network modules that contain immune response genes associated with T1D, we observed highly dynamic differences in module connectivity in the 600 days (~ 2 years) preceding clinical diagnosis of T1D. Our results suggest that gene co-expression is highly plastic and that connectivity differences in T1D-associated immune system genes influence the timing and development of clinical disease.
BibTeX
@inproceedings{brnne2021dynamic,
author = {Ingrid Brænne and Suna Önengüt-Gümüşcü and Ruoxi Chen and Ani Manichaikul and Stephen S. Rich and Wei‐Min Chen and Charles R. Farber and Marian Rewers and Aaron Barbour and Kimberly Bautista and Judith Baxter and Daniel Felipe-Morales and Kimberly A. Driscoll and Brigitte I. Frohnert and Marisa Stahl and Patricia Gesualdo and Michelle Hoffman and Rachel Karban and Edwin Liu and Jill Norris and Stesha Peacock and Hanan Shorrosh and Andrea K. Steck and Megan Stern and Erica Villegas and Kathleen Waugh and Jorma Toppari and Olli Simell and Annika Adamsson and Suvi Ahonen and Iris Erlund and Leena Hakola and Anne Hekkala and Henna Holappa and Heikki Hyöty and Anni Ikonen and Jorma Ilonen and Sinikka Jäminki and Sanna Jokipuu and Leena Eklund Karlsson and Jukka Kero and Miia Kähönen and Mikael Knip and Minna-Liisa Koivikko and Merja Koskinen and Mirva Koreasalo and Kalle Kurppa and Jarita Kytölä and Tiina Latva-aho and Katri Lindfors and Maria Lönnrot and Elina Mäntymäki and Markus Mattila and Maija E. Miettinen and Katja Multasuo and Teija Mykkänen and Tiina Niininen and Sari Niinistö and Mia Nyblom and Sami Oikarinen and Paula Ollikainen and Zhian Othmani and Sirpa Pohjola and Petra Rajala and Jenna Rautanen and Anne Riikonen and Eija Riski and Miia Pekkola and Minna Romo and Satu Ruohonen and Satu Simell and Maija Sjöberg and Aino Stenius and Päivi Tossavainen and Mari Vähä-Mäkilä and Sini Vainionpää and Eeva Varjonen and Riitta Veijola and Irene Viinikangas and Suvi Μ. Virtanen and Jin-Xiong She and Desmond Schatz and Diane Hopkins and Leigh Steed and Jennifer Bryant and Katherine Silvis and Michael Haller and Melissa Gardiner and Richard McIndoe and Ashok Sharma and Stephen W. Anderson and Laura M. Jacobsen and John Marks and P.D. Towe and Anette G. Ziegler and Ezio Bonifacio and Anita Gavrisan and Cigdem Gezginci and Anja Heublein and Verena Hoffmann},
title = {{Dynamic changes in immune gene co-expression networks predict development of type 1 diabetes}},
booktitle = {Scientific},
year = {2021},
doi = {10.1038/s41598-021-01840-z},
}