Impact statement

Precision medicine is an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle. As defined by Christensen et al. ([2009], The Innovator’s Prescription: A Disruptive Solution for Health Care), precision medicine is provision of care for diseases that can be precisely diagnosed, whose causes are understood, and which consequently can be treated with rules-based therapies that are predictably effective. Autoimmune diseases are those in which the body’s natural defense system loses discriminating power between its own cells and foreign cells, causing the body to mistakenly attack healthy tissues. There are more than 80 types of autoimmune diseases that affect a wide range of organ systems. These conditions are very heterogeneous in their presentation and therefore difficult to diagnose and treat. Achieving precision medicine in autoimmune diseases has been challenging due to the complex etiologies of these conditions, involving an interplay between genetic, epigenetic, and environmental factors. However, recent technological and computational advances in molecular profiling have helped to identify patient subtypes and molecular pathways that can be used to improve diagnostics and therapeutics. This review discusses the current understanding of the disease mechanisms, heterogeneity, and pathogenic autoantigens in autoimmune diseases gained from genomic and transcriptomic studies and highlights how these findings can be applied to better understand disease heterogeneity. Within that framework, improved diagnostics and targeted therapeutic approaches may advance toward precision clinical care of patients with autoimmune diseases.

Resolving patient heterogeneity

Autoimmune diseases are frequently characterized by clinical features or autoantibody prevalence; however, these features are heterogeneous and often overlap between autoimmune diseases, hindering precise diagnosis, and early treatment. Therefore, moving toward molecular diagnostics, which define disease based on changes in biological molecules, may aid diagnosis, and improve clinical outcomes of autoimmune diseases. Recently, exome and genome sequencing have shown promise for identifying pathogenic genetic variants in cases of rare monogenic diseases (Boycott et al., Reference Boycott, Hartley, Biesecker, Gibbs, Innes, Riess, Belmont, Dunwoodie, Jojic, Lassmann, Mackay, Temple, Visel and Baynam2019), including patients with autoinflammatory diseases (Kosukcu et al., Reference Kosukcu, Taskiran, Batu, Sag, Bilginer, Alikasifoglu and Ozen2020) such as hereditary fever syndromes. However, these are the rare exceptions. The genetic causes of most autoimmune diseases are complex and genetic risk is determined predominantly by the human leukocyte antigens (HLA) locus, which has the strongest association to rheumatic diseases. Outside of the HLA region, which can account for up to 50% of the genetic risk of a given complex autoimmune trait, hundreds of variants identified through genome-wide association studies (GWASs) each have small additive individual effects, making a diagnosis of autoimmune diseases based solely on genetics currently impossible. In RA, the number of RA-associated risk alleles weighted by the odds ratio correlates with disease risk; however, the predictive power of genetic risk scores is modest and not currently suitable for use in clinical practice (Karlson et al., Reference Karlson, Chibnik, Kraft, Cui, Keenan, Ding, Raychaudhuri, Klareskog, Alfredsson and Plenge2010; Dudbridge, Reference Dudbridge2013).

As genetic variants identified by GWAS are common variants (generally found in 1% or more of the population – a consequence of study design) and only modestly increase the risk of autoimmune diseases, rare variants with strong effects may contribute to the missing heritability of some patients with autoimmune diseases. For example, following the discovery of mutations in the TREX1 gene causing the type I interferonopathy Aicardi–Goutières syndrome, TREX1 variants were identified in up to 0.5–2% of patients with SLE (Lee-Kirsch et al., Reference Lee-Kirsch, Gong, Chowdhury, Senenko, Engel, Lee, de Silva, Bailey, Witte, Vyse, Kere, Pfeiffer, Harvey, Wong, Koskenmies, Hummel, Rohde, Schmidt, Dominiczak, Gahr, Hollis, Perrino, Lieberman and Hübner2007; Namjou et al., Reference Namjou, Kothari, Kelly, Glenn, Ojwang, Adler, Alarcón-Riquelme, Gallant, Boackle, Criswell, Kimberly, Brown, Edberg, Stevens, Jacob, Tsao, Gilkeson, Kamen, Merrill, Petri, Goldman, Vila, Anaya, Niewold, Martin, Pons-Estel, Sabio, Callejas, Vyse, Bae, Perrino, Freedman, Scofield, Moser, Gaffney, James, Langefeld, Kaufman, Harley and Atkinson2011). More recently, exome sequencing identified two rare variants in BLK and BANK1 in a subset of patients with SLE that increased type I interferon (IFN) activity (Jiang et al., Reference Jiang, Athanasopoulos, Ellyard, Chuah, Cappello, Cook, Prabhu, Cardenas, Gu, Stanley, Roco, Papa, Yabas, Walters, Burgio, McKeon, Byers, Burrin, Enders, Miosge, Canete, Jelusic, Tasic, Lungu, Alexander, Kitching, Fulcher, Shen, Arsov, Gatenby, Babon, Mallon, de Lucas Collantes, Stone, Wu, Field, Andrews, Cho, Pascual, Cook and Vinuesa2019). A recently published paper illustrated the role of rare variants in TLR-7 in monogenic SLE demonstrating that with more accessible and available whole exome and genome sequencing, we will learn more about the role of rare variants in autoimmune diseases (Brown et al., Reference Brown, Cañete, Wang, Medhavy, Bones, Roco, He, Qin, Cappello, Ellyard, Bassett, Shen, Burgio, Zhang, Turnbull, Meng, Wu, Cho, Miosge, Andrews, Field, Tvorogov, Lopez, Babon, López, Gónzalez-Murillo, Garulo, Pascual, Levy, Mallack, Calame, Lotze, Lupski, Ding, Ullah, Walters, Koina, Cook, Shen, de Lucas Collantes, Corry, Gantier, Athanasopoulos and Vinuesa2022). Together, these studies suggest that rare variants may contribute to the genetic risk and clinical heterogeneity of autoimmune diseases. However, the extensive heterogeneity within each autoimmune disease suggests that multiple pathways may contribute to disease; therefore, identifying subgroups of patients with shared molecular signatures is the best avenue to improve the diagnosis and treatment of patients with autoimmune diseases.

As one example, multiple studies have determined subsets of patients with SLE using transcriptomic approaches (Lyons et al., Reference Lyons, Rayner, Trivedi, Holle, Watts, Jayne, Baslund, Brenchley, Bruchfeld, Chaudhry, Cohen Tervaert, Deloukas, Feighery, Gross, Guillevin, Gunnarsson, Harper, Hrušková, Little, Martorana, Neumann, Ohlsson, Padmanabhan, Pusey, Salama, Sanders, Savage, Segelmark, Stegeman, Tesař, Vaglio, Wieczorek, Wilde, Zwerina, Rees, Clayton and Smith2012; Banchereau et al., Reference Banchereau, Hong, Cantarel, Baldwin, Baisch, Edens, Cepika, Acs, Turner, Anguiano, Vinod, Kahn, Obermoser, Blankenship, Wakeland, Nassi, Gotte, Punaro, Liu, Banchereau, Rossello-Urgell, Wright and Pascual2016; Toro-Domínguez et al., Reference Toro‐Domínguez, Martorell‐Marugán, Goldman, Petri, Carmona‐Sáez and Alarcón‐Riquelme2018; Figgett et al., Reference Figgett, Monaghan, Ng, Alhamdoosh, Maraskovsky, Wilson, Hoi, Morand and Mackay2019; Panousis et al., Reference Panousis, Bertsias, Ongen, Gergianaki, Tektonidou, Trachana, Romano-Palumbo, Bielser, Howald, Pamfil, Fanouriakis, Kosmara, Repa, Sidiropoulos, Dermitzakis and Boumpas2019; Andreoletti et al., Reference Andreoletti, Lanata, Trupin, Paranjpe, Jain, Nititham, Taylor, Combes, Maliskova, Ye, Katz, Dall’Era, Yazdany, Criswell and Sirota2021; Sandling et al., Reference Sandling, Pucholt, Hultin Rosenberg, Farias, Kozyrev, Eloranta, Alexsson, Bianchi, Padyukov, Bengtsson, Jonsson, Omdal, Lie, Massarenti, Steffensen, Jakobsen, Lillevang, Lerang, Molberg, Voss, Troldborg, Jacobsen, Syvänen, Jönsen, Gunnarsson, Svenungsson, Rantapää-Dahlqvist, Bengtsson, Sjöwall, Leonard, Lindblad-Toh and Rönnblom2021). Initial investigations found that approximately half of the patients with SLE exhibit increased peripheral blood expression of type I IFN-regulated genes, termed the “IFN signature,” associated with more severe disease (Baechler et al., Reference Baechler, Batliwalla, Karypis, Gaffney, Ortmann, Espe, Shark, Grande, Hughes, Kapur, Gregersen and Behrens2003; Bennett et al., Reference Bennett, Palucka, Arce, Cantrell, Borvak, Banchereau and Pascual2003), suggesting that a subset of patients with SLE may benefit from therapies targeting the IFN pathway. Consistent with these findings, the recently approved monoclonal antibody anifrolumab, which targets the type I IFN receptor subunit 1, is effective in about 16% of patients with SLE (Morand et al., Reference Morand, Furie, Tanaka, Bruce, Askanase, Richez, Bae, Brohawn, Pineda, Berglind and Tummala2020). However, there are conflicting reports regarding the effectiveness of stratifying patients based on IFN gene signatures in clinical trials of type I IFN inhibition (Khamashta et al., Reference Khamashta, Merrill, Werth, Furie, Kalunian, Illei, Drappa, Wang and Greth2016; Furie et al., Reference Furie, Khamashta, Merrill, Werth, Kalunian, Brohawn, Illei, Drappa, Wang and Yoo2017; Morand et al., Reference Morand, Furie, Tanaka, Bruce, Askanase, Richez, Bae, Brohawn, Pineda, Berglind and Tummala2020), demonstrating the complexity of the type I IFN response in SLE and identifying the need for additional stratification approaches.

To further refine the IFN signature in SLE, Chiche et al. (Reference Chiche, Jourde‐Chiche, Whalen, Presnell, Gersuk, Dang, Anguiano, Quinn, Burtey, Berland, Kaplanski, Harle, Pascual and Chaussabel2014) found that three distinct transcriptional IFN groups or modules were associated with 87% of patients with SLE and that all types of IFN, not just type I IFN, contributed to the IFN signatures. Importantly, patients with SLE could be further stratified based on the number of active IFN modules (Chiche et al., Reference Chiche, Jourde‐Chiche, Whalen, Presnell, Gersuk, Dang, Anguiano, Quinn, Burtey, Berland, Kaplanski, Harle, Pascual and Chaussabel2014). In 2016, Banchereau et al. (Reference Banchereau, Hong, Cantarel, Baldwin, Baisch, Edens, Cepika, Acs, Turner, Anguiano, Vinod, Kahn, Obermoser, Blankenship, Wakeland, Nassi, Gotte, Punaro, Liu, Banchereau, Rossello-Urgell, Wright and Pascual2016) confirmed and extended these findings in a large cohort of pediatric patients with SLE, identifying overexpression of additional transcriptional modules that correlated with disease activity and clinical parameters of SLE (Banchereau et al., Reference Banchereau, Hong, Cantarel, Baldwin, Baisch, Edens, Cepika, Acs, Turner, Anguiano, Vinod, Kahn, Obermoser, Blankenship, Wakeland, Nassi, Gotte, Punaro, Liu, Banchereau, Rossello-Urgell, Wright and Pascual2016). In addition, patients were stratified into seven clusters based on five immune signatures correlating with disease activity, including type I IFN-, neutrophil-, and plasmablast-associated signatures (Banchereau et al., Reference Banchereau, Hong, Cantarel, Baldwin, Baisch, Edens, Cepika, Acs, Turner, Anguiano, Vinod, Kahn, Obermoser, Blankenship, Wakeland, Nassi, Gotte, Punaro, Liu, Banchereau, Rossello-Urgell, Wright and Pascual2016). Using a similar approach, Toro‐Domínguez et al. (Reference Toro‐Domínguez, Martorell‐Marugán, Goldman, Petri, Carmona‐Sáez and Alarcón‐Riquelme2018, Reference Toro-Domínguez, Lopez-Domínguez, García Moreno, Villatoro-García, Martorell-Marugán, Goldman, Petri, Wojdyla, Pons-Estel, Isenberg, Morales-Montes de Oca, Trejo-Zambrano, García González, Rosetti, Gómez-Martín, Romero-Díaz, Carmona-Sáez and Alarcón-Riquelme2019) identified three SLE patient clusters characterized by a lymphocyte or neutrophil signature that may respond differently to treatments.

Most transcriptomic studies in SLE use whole blood or bulk cell input making it difficult to discern the affected cell populations. Therefore, single-cell analyses may be necessary to identify and refine molecular clusters in disease-relevant cell state (Perez et al., Reference Papp, Gordon, Thaçi, Morita, Gooderham, Foley, Girgis, Kundu and Banerjee2022). Using single-cell RNA sequencing, Nehar-Belaid et al. (Reference Papp, Gordon, Thaçi, Morita, Gooderham, Foley, Girgis, Kundu and Banerjee2020) defined the cellular subgroups that contributed to the IFN signature in pediatric SLE, including T cells, dendritic cells (DCs), monocytes, and natural killer (NK) cells. Notably, the clustering of these cell types revealed six distinct subgroups of patients associated with disease activity (Nehar-Belaid et al., Reference Papp, Gordon, Thaçi, Morita, Gooderham, Foley, Girgis, Kundu and Banerjee2020). In a recent study, Andreoletti et al. (Reference Andreoletti, Lanata, Trupin, Paranjpe, Jain, Nititham, Taylor, Combes, Maliskova, Ye, Katz, Dall’Era, Yazdany, Criswell and Sirota2021) determined unique subgroups of patients based on the transcriptional profiles of sorted monocytes, B cells, CD4+ T cells, and NK cells that correlated with disease activity and ethnicity. In addition, multi-omic approaches may also improve patient stratification, as seen in a study by Guthridge et al. (Reference Guthridge, Lu, Tran, Arriens, Aberle, Kamp, Munroe, Dominguez, Gross, DeJager, Macwana, Bourn, Apel, Thanou, Chen, Chakravarty, Merrill and James2020), in which integration of transcriptional modules and autoantibody and soluble mediator profiles identified seven patient clusters with distinct molecular pathways but similar clinical outcomes. In another study, Lanata et al. (Reference Lanata, Paranjpe, Nititham, Taylor, Gianfrancesco, Paranjpe, Andrews, Chung, Rhead, Barcellos, Trupin, Katz, Dall’Era, Yazdany, Sirota and Criswell2019) used clinical features to define three distinct subgroups of SLE with unsupervised clustering that was supported by differential methylation patterns and ethnicity. Several of these studies explore multi-ethnic cohorts. There are known differences in SLE disease manifestations and severity across different racial and ethnic groups. When exploring biological differences across different patient groups, it’s important to note the potential inaccuracy or lack of specification between self-reported and genetic-driven subgroups which may contribute to interpretation problems as ethnicity may be more predictive of differences due to disparities than genetic background (Mersha and Abebe, Reference Mersha and Abebe2015).

Interestingly, genomic studies have found that autoimmune diseases have shared genetic associations, suggesting that similar pathogenic mechanisms may contribute to different autoimmune diseases (Zhernakova et al., Reference Zhernakova, van Diemen and Wijmenga2009; Richard-Miceli and Criswell, Reference Richard-Miceli and Criswell2012). Indeed, transcriptome and methylome analysis of patients with seven autoimmune diseases demonstrated four patient clusters that differed in the expression of inflammatory, lymphoid, or IFN signature (Barturen et al., Reference Barturen, Babaei, Català-Moll, Martínez-Bueno, Makowska, Martorell-Marugán, Carmona-Sáez, Toro-Domínguez, Carnero-Montoro, Teruel, Kerick, Acosta-Herrera, le Lann, Jamin, Rodríguez-Ubreva, García-Gómez, Kageyama, Buttgereit, Hayat, Mueller, Lesche, Hernandez-Fuentes, Juarez, Rowley, White, Marañón, Gomes Anjos, Varela, Aguilar-Quesada, Garrancho, López-Berrio, Rodriguez Maresca, Navarro-Linares, Almeida, Azevedo, Brandão, Campar, Faria, Farinha, Marinho, Neves, Tavares, Vasconcelos, Trombetta, Montanelli, Vigone, Alvarez-Errico, Li, Thiagaran, Blanco Alonso, Corrales Martínez, Genre, López Mejías, Gonzalez-Gay, Remuzgo, Ubilla Garcia, Cervera, Espinosa, Rodríguez-Pintó, de Langhe, Cremer, Lories, Belz, Hunzelmann, Baerlecken, Kniesch, Witte, Lehner, Stummvoll, Zauner, Aguirre-Zamorano, Barbarroja, Castro-Villegas, Collantes-Estevez, Ramon, Díaz Quintero, Escudero-Contreras, Fernández Roldán, Jiménez Gómez, Jiménez Moleón, Lopez-Pedrera, Ortega-Castro, Ortego, Raya, Artusi, Gerosa, Meroni, Schioppo, de Groof, Ducreux, Lauwerys, Maudoux, Cornec, Devauchelle-Pensec, Jousse-Joulin, Jouve, Rouvière, Saraux, Simon, Alvarez, Chizzolini, Dufour, Wynar, Balog, Bocskai, Deák, Dulic, Kádár, Kovács, Cheng, Gerl, Hiepe, Khodadadi, Thiel, Rinaldis, Rao, Benschop, Chamberlain, Dow, Ioannou, Laigle, Marovac, Wojcik, Renaudineau, Borghi, Frostegård, Martín, Beretta, Ballestar, McDonald, Pers and Alarcón-Riquelme2021). Notably, patients with different autoimmune diseases were found within each cluster (Barturen et al., Reference Barturen, Babaei, Català-Moll, Martínez-Bueno, Makowska, Martorell-Marugán, Carmona-Sáez, Toro-Domínguez, Carnero-Montoro, Teruel, Kerick, Acosta-Herrera, le Lann, Jamin, Rodríguez-Ubreva, García-Gómez, Kageyama, Buttgereit, Hayat, Mueller, Lesche, Hernandez-Fuentes, Juarez, Rowley, White, Marañón, Gomes Anjos, Varela, Aguilar-Quesada, Garrancho, López-Berrio, Rodriguez Maresca, Navarro-Linares, Almeida, Azevedo, Brandão, Campar, Faria, Farinha, Marinho, Neves, Tavares, Vasconcelos, Trombetta, Montanelli, Vigone, Alvarez-Errico, Li, Thiagaran, Blanco Alonso, Corrales Martínez, Genre, López Mejías, Gonzalez-Gay, Remuzgo, Ubilla Garcia, Cervera, Espinosa, Rodríguez-Pintó, de Langhe, Cremer, Lories, Belz, Hunzelmann, Baerlecken, Kniesch, Witte, Lehner, Stummvoll, Zauner, Aguirre-Zamorano, Barbarroja, Castro-Villegas, Collantes-Estevez, Ramon, Díaz Quintero, Escudero-Contreras, Fernández Roldán, Jiménez Gómez, Jiménez Moleón, Lopez-Pedrera, Ortega-Castro, Ortego, Raya, Artusi, Gerosa, Meroni, Schioppo, de Groof, Ducreux, Lauwerys, Maudoux, Cornec, Devauchelle-Pensec, Jousse-Joulin, Jouve, Rouvière, Saraux, Simon, Alvarez, Chizzolini, Dufour, Wynar, Balog, Bocskai, Deák, Dulic, Kádár, Kovács, Cheng, Gerl, Hiepe, Khodadadi, Thiel, Rinaldis, Rao, Benschop, Chamberlain, Dow, Ioannou, Laigle, Marovac, Wojcik, Renaudineau, Borghi, Frostegård, Martín, Beretta, Ballestar, McDonald, Pers and Alarcón-Riquelme2021). Studies using immunophenotyping (Kroef et al., Reference Kroef, Hoogen, Mertens, SLM, Haskett, Devaprasad, Carvalheiro, Chouri, Vazirpanah, Cossu, CGK, Silva-Cardoso, Affandi, CPJ, Lopes, Hillen, Bonte-Mineur, Kok, Beretta, Rossato, Mingueneau, van Roon and Radstake2020; Martin‐Gutierrez et al., Reference Martin-Gutierrez, Peng, Thompson, Robinson, Naja, Peckham, Wu, J’bari, Ahwireng, Waddington, Bradford, Varnier, Gandhi, Radmore, Gupta, Isenberg, Jury and Ciurtin2021) and soluble mediator profiling (Slight-Webb et al., Reference Slight-Webb, Guthridge, Kheir, Chen, Tran, Gross, Roberts, Khan, Peercy, Saunkeah, Guthridge and James2021) also found that patients with different autoimmune diseases share similar molecular signatures. Thus, diagnosing patients based on molecular signatures in addition to clinical features may be a key step in moving toward precision medicine and targeted therapeutics.

Taken together, it becomes clear that autoimmune disorders comprise a wide spectrum of clinical manifestations. With the use of genomics, transcriptomics, and other multi-omic approaches, we can begin to examine these complex disorders under a magnifying glass to better define patient heterogeneity and identify targetable genes and pathways.

Defining the Autoantigenome

As autoimmune disorders are characterized by the body’s response to self, defining that exact “self” is critical to both treatment and diagnosis. Autoantibodies, a key component of disease that often directly contribute to outcomes, provide a window into defining these self-antigens and peptides. Of course, autoantibodies do not develop in a vacuum, and certain HLA alleles are strongly associated with autoimmune diseases (Liu et al., Reference Liu, Shao and Fu2021), indicating a key role for T cell help and antigen presentation in disease pathogenesis. Understanding and defining the interaction of these three components – autoantibodies, HLA alleles, and T cell repertoire – could identify novel therapeutic targets and molecular diagnostics.

The antibody and T cell repertoires are highly diverse due to recombination of variable, diversity, and joining gene segments, followed by somatic hypermutation in B cell receptors, making the identification of antigen specificity challenging. However, recent advances in Next Generation Sequencing (NGS) and computational approaches have enabled large-scale sequencing of antibody and T Cell Receptor (TCR) repertoires in autoimmune diseases (Zemlin et al., Reference Zemlin, Schelonka, Bauer and Schroeder2002; Schatz and Ji, Reference Tang, Wan, Wang, Pan, Wang and Chen2011; Rechavi and Somech, Reference Rechavi and Somech2017; Nielsen and Boyd, Reference Nehar-Belaid, Hong, Marches, Chen, Bolisetty, Baisch, Walters, Punaro, Rossi, Chung, Huynh, Singh, Flynn, Tabanor-Gayle, Kuchipudi, Mejias, Collet, Lucido, Palucka, Robson, Lakshminarayanan, Ramilo, Wright, Pascual and Banchereau2019; Nielsen et al., Reference Nielsen, Roskin, Jackson, Joshi, Nejad, Lee, Wagar, Pham, Hoh, Nguyen, Tsunemoto, Patel, Tibshirani, Ley, Davis, Parsonnet and Boyd2019).

Anti-citrullinated protein antibodies (ACPAs) that recognize the posttranslational modification of the amino acid citrulline are a hallmark of RA and contribute to disease pathogenesis (Kurowska et al., Reference Kurowska, Kuca-Warnawin, Radzikowska and Maśliński2017). Antibodies consist of two heavy- and light-chain pairs, which both contain antigen-binding domains; therefore, pairing heavy- and light-chains is necessary to determine antigen specificity (Robinson, Reference Robinson2015). To accomplish this, Tan et al. (Reference Zhang, Wei, Slowikowski, Fonseka, Rao, Kelly, Goodman, Tabechian, Hughes, Salomon-Escoto, Watts, Jonsson, Rangel-Moreno, Meednu, Rozo, Apruzzese, Eisenhaure, Lieb, Boyle, Mandelin, Boyce, DiCarlo, Gravallese, Gregersen, Moreland, Firestein, Hacohen, Nusbaum, Lederer, Perlman, Pitzalis, Filer, Holers, Bykerk, Donlin, Anolik, Brenner and Raychaudhuri2014) developed a novel DNA barcoding method to sequence heavy- and light-chain pairs from antibody-producing plasmablasts in ACPA-positive patients with RA and determined affinity-matured clonal families of antibodies. Recombinant expression of 14 antibodies identified four ACPAs with differential targeting of α-enolase, citrullinated fibrinogen, and citrullinated histone H2B (Tan et al., Reference Zhang, Wei, Slowikowski, Fonseka, Rao, Kelly, Goodman, Tabechian, Hughes, Salomon-Escoto, Watts, Jonsson, Rangel-Moreno, Meednu, Rozo, Apruzzese, Eisenhaure, Lieb, Boyle, Mandelin, Boyce, DiCarlo, Gravallese, Gregersen, Moreland, Firestein, Hacohen, Nusbaum, Lederer, Perlman, Pitzalis, Filer, Holers, Bykerk, Donlin, Anolik, Brenner and Raychaudhuri2014). Additional studies confirmed that ACPAs undergo affinity maturation, resulting in epitope spreading and polyreactivity with other post-translationally modified proteins (Elliott et al., Reference Elliott, Kongpachith, Lingampalli, Adamska, Cannon, Mao, Blum and Robinson2018; Titcombe et al., Reference Titcombe, Wigerblad, Sippl, Zhang, Shmagel, Sahlström, Zhang, Barsness, Ghodke-Puranik, Baharpoor, Hansson, Israelsson, Skriner, Niewold, Klareskog, Svensson, Amara, Malmström and Mueller2018; Kongpachith et al., Reference Kongpachith, Lingampalli, Ju, Blum, Lu, Elliott, Mao and Robinson2019; Steen et al., Reference Steen, Forsström, Sahlström, Odowd, Israelsson, Krishnamurthy, Badreh, Mathsson Alm, Compson, Ramsköld, Ndlovu, Rapecki, Hansson, Titcombe, Bang, Mueller, Catrina, Grönwall, Skriner, Nilsson, Lightwood, Klareskog and Malmström2019).

Repertoire analyses of plasmablasts from healthy individuals with RA-associated autoantibodies demonstrated elevated IgA responses (Kinslow et al., Reference Kinslow, Blum, Deane, Demoruelle, Okamoto, Parish, Kongpachith, Lahey, Norris, Robinson and Holers2016), suggesting that ACPAs may originate from mucosal immune responses. Furthermore, serial analyses of patients with RA found that ACPAs that persisted over time were predominantly IgA (Elliott et al., Reference Elliott, Kongpachith, Lingampalli, Adamska, Cannon, Mao, Blum and Robinson2018), consistent with continued mucosal antigen exposure. Therefore, identifying the specific mucosal antigens targeted by these ACPAs may help identify tolerizing therapies for patients with RA.

Early studies have identified expanded CD4+ T cell clones in the peripheral blood and synovial tissue of patients with RA (Goronzy et al., Reference Goronzy, Bartz-Bazzanella, Hu, Jendro, Walser-Kuntz and Weyand1994; Ikeda et al., Reference Ikeda, Masuko, Nakai, Kato, Hasunuma, Mizushima, Nishioka, Yamamoto and Yoshino1996; Schmidt et al., Reference Schmidt, Martens, Weyand and Goronzy1996; VanderBorght et al., Reference VanderBorght, Geusens, Vandevyver, Raus and Stinissen2000; Wagner et al., Reference Wagner, Pierer, Kaltenhäuser, Wilke, Seidel, Arnold and Häntzschel2003), including early in the disease course (Klarenbeek et al., Reference Klarenbeek, de Hair, Doorenspleet, van Schaik, Esveldt, van de Sande, Cantaert, Gerlag, Baeten, van Kampen, Baas, Tak and de Vries2012). Phenotypic analysis combining TCR sequencing and single-cell transcriptomics revealed expanded memory CD4+ T cell clones with upregulated senescence-related transcripts, chemokine receptors, and CD5 expression, suggestive of antigen stimulation and autoreactivity (Ishigaki et al., Reference Ishigaki, Shoda, Kochi, Yasui, Kadono, Tanaka, Fujio and Yamamoto2015). However, the autoantigens targeted by CD4+ T cells in RA remain elusive.

The HLA-DRB1 RA susceptibility alleles contain five shared amino acids of the β1 subunit, referred to as the shared epitope, which is associated with ACPA production (van Gaalen et al., Reference van Gaalen, van Aken, Huizinga, Schreuder, Breedveld, Zanelli, van Venrooij, Verweij, Toes and de Vries2004; Huizinga et al., Reference Huizinga, Amos, van der Helm-van Mil, Chen, van Gaalen, Jawaheer, Schreuder, Wener, Breedveld, Ahmad, Lum, de Vries, Gregersen, Toes and Criswell2005; Busch et al., Reference Busch, Kollnberger and Mellins2019). There is also significant clinical evidence of differential response based on mechanism in RA patients based on their ACPA/HLA epitope. HLA-DRB1 risk alleles for RA are associated with differential clinical responsiveness to abatacept and adalimumab according to the data from a head-to-head, randomized, single-blind study in autoantibody-positive early RA (Rigby et al., Reference Rigby, Buckner, Louis Bridges, Nys, Gao, Polinsky, Ray and Bykerk2021). GWAS analysis demonstrated that an amino acid within the P4 pocket of the peptide-binding groove strongly contributed to the association of HLA-DRB1 and RA (Raychaudhuri et al., Reference Raychaudhuri, Sandor, Stahl, Freudenberg, Lee, Jia, Alfredsson, Padyukov, Klareskog, Worthington, Siminovitch, Bae, Plenge, Gregersen and de Bakker2012), suggesting that the shared epitope may allow binding and presentation of citrullinated autoantigens. Consistent with this hypothesis, antigen discovery analyses using peptide stimulation or peptide–MHC tetramers revealed Th1 and Th17 reactivity to citrullinated antigens, including α-enolase, fibrinogen, vimentin, and aggrecan, in the peripheral blood of patients with RA (Delwig et al., Reference Delwig, Locke, Robinson and Ng2010; Law et al., Reference Law, Street, Yu, Capini, Ramnoruth, Nel, van Gorp, Hyde, Lau, Pahau, Purcell and Thomas2012; Scally et al., Reference Scally, Petersen, Law, Dudek, Nel, Loh, Wijeyewickrema, Eckle, van Heemst, Pike, McCluskey, Toes, la Gruta, Purcell, Reid, Thomas and Rossjohn2013; James et al., Reference James, Rieck, Pieper, Gebe, Yue, Tatum, Peda, Sandin, Klareskog, Malmström and Buckner2014; Gerstner et al., Reference Gerstner, Turcinov, Hensvold, Chemin, Uchtenhagen, Ramwadhdoebe, Dubnovitsky, Kozhukh, Rönnblom, Kwok, Achour, Catrina, van Baarsen and Malmström2020). In addition, T cells specific for citrullinated fibrinogen contribute to the development and progression of RA in mouse models (Hill et al., Reference Hill, Bell, Brintnell, Yue, Wehrli, Jevnikar, Lee, Hueber, Robinson and Cairns2008; Cordova et al., Reference Cordova, Willis, Haskins and Holers2013).

Although progress has been made in the identification of autoantigens targeted in autoimmune diseases using microarrays, mass spectrometry, and phage-display assays, these approaches are limited by the need to prespecify the antigens to be studied. Therefore, due to the large number and diversity of antibodies and TCRs, computational methods are needed to predict target antigens from the TCR or antibody sequence alone. Recent progress has been made to predict TCR specificity based on the hypothesis that TCRs that recognize the same antigen share CDR3 sequence motifs. In two separate studies, Dash et al. (Reference Dash, Fiore-Gartland, Hertz, Wang, Sharma, Souquette, Crawford, Clemens, Nguyen, Kedzierska, La Gruta, Bradley and Thomas2017) and Glanville et al. (Reference Glanville, Huang, Nau, Hatton, Wagar, Rubelt, Ji, Han, Krams, Pettus, Haas, CSL, Sette, Boyd, Scriba, Martinez and Davis2017) developed different algorithms (TCRdist – https://tcrdist3.readthedocs.io/en/latest/ and GLIPH – http://50.255.35.37:8080/, respectively) that clustered TCRs dependent on CDR3 motifs and accurately defined TCR specificity based on these clusters. However, although these approaches are promising, they are limited by the availability of pre-existing knowledge of TCR specificities to make predictions, and large-scale approaches to define these interactions are required. In a recent study, Zhang et al. (Reference Zhang, Liu, Zhang, Chen, Ye, Shukla, Qiao, Zhan, Chen, Wu, Fu and Li2020) clustered tumor TCRs based on antigen-specificity using iSMART and identified novel antigens by integrating TCR clusters, tumor genomics, and HLA genotypes (Zhang et al., Reference Zhang, Liu, Zhang, Chen, Ye, Shukla, Qiao, Zhan, Chen, Wu, Fu and Li2020). Therefore, multi-omic approaches paired with CDR3 clustering may also help define novel antigens targeted in autoimmune diseases.

In terms of precision medicine, a better understanding of the antigens, TCRs, HLAs, and BCRs driving disease offers a therapeutic window into these diverse disorders. Tolerizing therapies that target specific peptides or regulatory CAR-T cells offer a way to directly suppress autoimmune responses on a patient-by-patient basis.

The path to targeted therapeutics

Recent genomic and transcriptomic approaches have determined novel pathogenic mechanisms and begun to unravel the heterogeneity of autoimmune diseases, revealing potential therapeutic targets for precision medicine. This section will discuss current work applying knowledge obtained through genomic and transcriptomic studies toward precision medicine approaches.

Discovering novel therapeutic targets

Genetic analyses of monogenic autoinflammatory diseases have been pivotal in identifying druggable targets that are now used in clinical care (Manthiram et al., Reference Manthiram, Zhou, Aksentijevich and Kastner2017). For example, therapies targeting IL-1, such as anakinra, are approved for the inflammasomopathy cryopyrin-associated periodic fever syndrome (Hoffman, Reference Hoffman2009). Genetic studies have also revealed efficacious therapeutic targets in polygenic autoimmune disorders. Genetic variation in the Janus kinase family member tyrosine kinase 2 (TYK2), required for type 1 IFN, IL-12 and IL-23 signaling (Sohn et al., Reference Sohn, Barrett, van Abbema, Chang, Kohli, Kanda, Smith, Lai, Zhou, Zhang, Yang, Williams, Macleod, Hurley, Kulagowski, Lewin-Koh, Dengler, Johnson, Ghilardi, Zak, Liang, Blair, Magnuson and Wu2013; Burke et al., Reference Burke, Cheng, Gillooly, Strnad, Zupa-Fernandez, Catlett, Zhang, Heimrich, Mcintyre, Cunningham, Carman, Zhou, Banas, Chaudhry, Li, D’Arienzo, Chimalakonda, Yang, Xie, Pang, Zhao, Rose, Huang, Moslin, Wrobleski, Weinstein and Salter-Cid2019), is associated with autoimmune diseases, including psoriasis (Genetic Analysis of Psoriasis Consortium & the Wellcome Trust Case Control Consortium Reference Andreoletti, Lanata, Trupin, Paranjpe, Jain, Nititham, Taylor, Combes, Maliskova, Ye, Katz, Dall’Era, Yazdany, Criswell and Sirota2 et al., Reference Strange, Capon, Spencer, Knight, Weale, Allen, Barton, Band, Bellenguez, Bergboer, Blackwell, Bramon, Bumpstead, Casas, Cork, Corvin, Deloukas, Dilthey, Duncanson, Edkins, Estivill, Fitzgerald, Freeman, Giardina, Gray, Hofer, Hüffmeier, Hunt, Irvine, Jankowski, Kirby, Langford, Lascorz, Leman, Leslie, Mallbris, Markus, Mathew, McLean, McManus, Mössner, Moutsianas, Naluai, Nestle, Novelli, Onoufriadis, Palmer, Perricone, Pirinen, Plomin, Potter, Pujol, Rautanen, Riveira-Munoz, Ryan, Salmhofer, Samuelsson, Sawcer, Schalkwijk, Smith, Ståhle, Su, Tazi-Ahnini, Traupe, Viswanathan, Warren, Weger, Wolk, Wood, Worthington, Young, Zeeuwen, Hayday, Burden, Griffiths, Kere, Reis, McVean, Evans, Brown, Barker, Peltonen, Donnelly and Trembath2010; Ellinghaus et al., Reference Ellinghaus, Ellinghaus, Nair, Stuart, Esko, Metspalu, Debrus, Raelson, Tejasvi, Belouchi, West, Barker, Kõks, Kingo, Balschun, Palmieri, Annese, Gieger, Wichmann, Kabesch, Trembath, Mathew, Abecasis, Weidinger, Nikolaus, Schreiber, Elder, Weichenthal, Nothnagel and Franke2012; Tsoi et al., Reference Tsoi, Spain, Knight, Ellinghaus, Stuart, Capon, Ding, Li, Tejasvi, Gudjonsson, Kang, Allen, McManus, Novelli, Samuelsson, Schalkwijk, Ståhle, Burden, Smith, Cork, Estivill, Bowcock, Krueger, Weger, Worthington, Tazi-Ahnini, Nestle, Hayday, Hoffmann, Winkelmann, Wijmenga, Langford, Edkins, Andrews, Blackburn, Strange, Band, Pearson, Vukcevic, Spencer, Deloukas, Mrowietz, Schreiber, Weidinger, Koks, Kingo, Esko, Metspalu, Lim, Voorhees, Weichenthal, Wichmann, Chandran, Rosen, Rahman, Gladman, Griffiths, Reis, Kere, Nair, Franke, Barker, Abecasis, Elder and Trembath2012), psoriatic arthritis (Mease et al., Reference Morand, Furie, Tanaka, Bruce, Askanase, Richez, Bae, Brohawn, Pineda, Berglind and Tummala2022), Crohn’s disease (Franke et al., Reference Franke, McGovern, Barrett, Wang, Radford-Smith, Ahmad, Lees, Balschun, Lee, Roberts, Anderson, Bis, Bumpstead, Ellinghaus, Festen, Georges, Green, Haritunians, Jostins, Latiano, Mathew, Montgomery, Prescott, Raychaudhuri, Rotter, Schumm, Sharma, Simms, Taylor, Whiteman, Wijmenga, Baldassano, Barclay, Bayless, Brand, Büning, Cohen, Colombel, Cottone, Stronati, Denson, de Vos, D’Inca, Dubinsky, Edwards, Florin, Franchimont, Gearry, Glas, van Gossum, Guthery, Halfvarson, Verspaget, Hugot, Karban, Laukens, Lawrance, Lemann, Levine, Libioulle, Louis, Mowat, Newman, Panés, Phillips, Proctor, Regueiro, Russell, Rutgeerts, Sanderson, Sans, Seibold, Steinhart, Stokkers, Torkvist, Kullak-Ublick, Wilson, Walters, Targan, Brant, Rioux, D’Amato, Weersma, Kugathasan, Griffiths, Mansfield, Vermeire, Duerr, Silverberg, Satsangi, Schreiber, Cho, Annese, Hakonarson, Daly and Parkes2010), and SLE (Sigurdsson et al., Reference Sigurdsson, Nordmark, Göring, Lindroos, Wiman, Sturfelt, Jönsen, Rantapää-Dahlqvist, Möller, Kere, Koskenmies, Widén, Eloranta, Julkunen, Kristjansdottir, Steinsson, Alm, Rönnblom and Syvänen2005; Graham et al., Reference Graham, Morris, Bhangale, Criswell, Syvänen, Rönnblom, Behrens, Graham and Vyse2011; Tang et al., Reference Tang, Wan, Wang, Pan, Wang and Chen2015; Lee and Bae, Reference Lee and Bae2016). In phase II and III clinical trials, the TYK2 inhibitor deucravacitinib (BMS-986165) was more effective compared to placebo in patients with moderate-to-severe plaque psoriasis (Papp et al., Reference Rechavi and Somech2018), and is now approved in the US, EU, and other regions. In addition, deucravacitinib is being investigated in early trials of Crohn’s disease and SLE demonstrating efficacy in phase II trials in PsA (Mease et al., Reference Morand, Furie, Tanaka, Bruce, Askanase, Richez, Bae, Brohawn, Pineda, Berglind and Tummala2022) and SLE (Morand et al., Reference Morand, Pike, Merrill, van Vollenhoven, Werth, Hobar, Delev, Shah, Sharkey, Wegman, Catlett, Banerjee and Singhal2023). However, with polygenic autoimmune diseases, not all identified gene variants may be effective drug targets. Therefore, moving beyond individual genes toward gene networks using in silico drug efficacy screening, such as drug–disease network proximity analyses (Kim et al., Reference Kim, Moon, Park and Tagkopoulos2020), to predict potential therapies is needed for drug discovery in autoimmunity. Using this approach, Cordell et al. (Reference Cordell, Fryett, Ueno, Darlay, Aiba, Hitomi, Kawashima, Nishida, Khor, Gervais, Kawai, Nagasaki, Tokunaga, Tang, Shi, Li, Juran, Atkinson, Gerussi, Carbone, Asselta, Cheung, de Andrade, Baras, Horowitz, MAR, Sun, Jones, Flack, Spicer, Mulcahy, Byun, Han, Sandford, Lazaridis, Amos, Hirschfield, Seldin, Invernizzi, Siminovitch, Ma, Nakamura and Mells2021) recently identified 56 genetic variants associated with primary biliary cholangitis in a genome-wide meta-analysis and predicted several candidate therapies for the disease, including approved treatments of other autoimmune diseases.

Translating genomics to cell function may also identify potentially targetable pathways. Smillie et al. (Reference Smillie, Biton, Ordovas-Montanes, Sullivan, Burgin, Graham, Herbst, Rogel, Slyper, Waldman, Sud, Andrews, Velonias, Haber, Jagadeesh, Vickovic, Yao, Stevens, Dionne, Nguyen, Villani, Hofree, Creasey, Huang, Rozenblatt-Rosen, Garber, Khalili, Desch, Daly, Ananthakrishnan, Shalek, Xavier and Regev2019) created a cell atlas of UC using single-cell transcriptomics, highlighting the cells that change in proportions or gene expression compared to healthy tissues. In addition, mapping UC-associated risk alleles onto the cell atlas demonstrated enrichment of risk alleles in individual cell lineages, including M-like cells that exhibited high expression of multiple risk alleles, providing important information about disease etiology and molecular pathways (Smillie et al., Reference Smillie, Biton, Ordovas-Montanes, Sullivan, Burgin, Graham, Herbst, Rogel, Slyper, Waldman, Sud, Andrews, Velonias, Haber, Jagadeesh, Vickovic, Yao, Stevens, Dionne, Nguyen, Villani, Hofree, Creasey, Huang, Rozenblatt-Rosen, Garber, Khalili, Desch, Daly, Ananthakrishnan, Shalek, Xavier and Regev2019). There are several other large-scale efforts to generate single-cell transcriptomic and proteomic datasets in RA and SLE as well as other autoimmune diseases that are able to elucidate cell type specific disease associated genes and pathways (Zhang et al., Reference Zhang, Wei, Slowikowski, Fonseka, Rao, Kelly, Goodman, Tabechian, Hughes, Salomon-Escoto, Watts, Jonsson, Rangel-Moreno, Meednu, Rozo, Apruzzese, Eisenhaure, Lieb, Boyle, Mandelin, Boyce, DiCarlo, Gravallese, Gregersen, Moreland, Firestein, Hacohen, Nusbaum, Lederer, Perlman, Pitzalis, Filer, Holers, Bykerk, Donlin, Anolik, Brenner and Raychaudhuri2019). However, comprehensive multi-disease cell atlases are needed to provide further insights, which require the integration of multiple large-scale studies and data sets that may have been collected under different conditions. To avoid confounding variables between studies, multiple computational approaches have been created to remove batch effects (Butler et al., Reference Butler, Hoffman, Smibert, Papalexi and Satija2018; Haghverdi et al., Reference Haghverdi, Lun, Morgan and Marioni2018; Hie et al., Reference Hie, Bryson and Berger2019; Korsunsky et al., Reference Korsunsky, Millard, Fan, Slowikowski, Zhang, Wei, Baglaenko, Brenner, Loh and Raychaudhuri2019; Polański et al., Reference Polański, Park, Young, Miao, Meyer and Teichmann2020; Tran et al., Reference Tran, Ang, Chevrier, Zhang, Lee, Goh and Chen2020). As one example, the algorithm Harmony (Korsunsky et al., Reference Korsunsky, Millard, Fan, Slowikowski, Zhang, Wei, Baglaenko, Brenner, Loh and Raychaudhuri2019) was used to integrate single-cell transcriptomic profiles from multiple disease datasets, revealing a CXCL10+CCL2+ inflammatory macrophage phenotype in the tissues of patients with RA, Crohn’s disease, UC, and COVID-19 (Consortium et al., Reference Zhang, Mears, Shakib, Beynor, Shanaj, Korsunsky, Nathan, Donlin and Raychaudhuri2021), suggesting that the same pathway may be targeted in distinct diseases.

Antigen-specific therapies

Identifying antigens targeted by antibodies and T cells in autoimmune diseases will allow for the development of antigen-specific therapies, aiming to restore immune tolerance in autoreactive lymphocytes while maintaining overall immune surveillance to infections and cancer. Tolerogenic DCs have been tested in early phase clinical trials for multiple autoimmune diseases, such as RA, Crohn’s disease, and MS (Phillips et al., Reference Phillips, Garciafigueroa, Trucco and Giannoukakis2017). In a recent phase 1b trial, autologous tolerogenic DCs loaded with myelin-derived antigens and aquaporin-4 were analyzed for efficacy in MS and neuromyelitis optica spectrum disorders (NMOSDs; Zubizarreta et al., Reference Zubizarreta, Flórez-Grau, Vila, Cabezón, España, Andorra, Saiz, Llufriu, Sepulveda, Sola-Valls, Martinez-Lapiscina, Pulido-Valdeolivas, Casanova, Martinez Gines, Tellez, Oreja-Guevara, Español, Trias, Cid, Juan, Lozano, Blanco, Steinman, Benitez-Ribas and Villoslada2019). The tolerogenic DC therapy was well-tolerated and induced IL-10 production by peptide-stimulated cells and a trend toward an increase in regulatory T cells, compatible with tolerance induction (Zubizarreta et al., Reference Zubizarreta, Flórez-Grau, Vila, Cabezón, España, Andorra, Saiz, Llufriu, Sepulveda, Sola-Valls, Martinez-Lapiscina, Pulido-Valdeolivas, Casanova, Martinez Gines, Tellez, Oreja-Guevara, Español, Trias, Cid, Juan, Lozano, Blanco, Steinman, Benitez-Ribas and Villoslada2019).

The therapeutic potential of polyclonal regulatory T cells has also been demonstrated in some autoimmune diseases, including MS (Kohm et al., Reference Kohm, Carpentier, Anger and Miller2002); however, the effects are mostly modest, possibly because of nonspecific regulatory T cells (Raffin et al., Reference Raffin, Vo and Bluestone2020). Therefore, based on knowledge acquired from T cell therapies in oncology, another approach is the generation of autologous antigen-specific regulatory T cells by transfecting TCRs or CARs for autoantigens. Kim et al. (Reference Kim, Zhang, Yoon, Culp, Lees, Wucherpfennig and Scott2018) transduced human regulatory T cells with a myelin-basic protein-specific TCR isolated from an MS patient and demonstrated that the MBP-specific regulatory T cells suppressed MBP-specific effector cells in vitro and ameliorated disease in a mouse model of MS. Therefore, although still in the preclinical phase, antigen-specific regulatory T cells show promise for the treatment of MS.

Computational drug repurposing

Identifying new uses for approved drugs, or drug repurposing, will also benefit precision medicine in autoimmune diseases, as conventional drug discovery is often costly and time consuming (DiMasi et al., Reference DiMasi, Hansen and Grabowski2003). However, traditional drug repurposing relies on high-throughput screening technologies that can also be costly; therefore, novel methods are required to expand drug repurposing efforts. Sirota et al. (Reference van Gaalen, van Aken, Huizinga, Schreuder, Breedveld, Zanelli, van Venrooij, Verweij, Toes and de Vries2011) developed a systematic computational approach to predict disease–drug relationships by comparing gene expression signatures of diseases with those of FDA-approved drugs. This approach identified a novel therapeutic association of an antiepileptic drug, topiramate, with inflammatory bowel disease that was efficacious in a rodent model of colitis (Dudley et al., Reference Dudley, Sirota, Shenoy, Pai, Roedder, Chiang, Morgan, Sarwal, Pasricha and Butte2011).

Caveats and conclusions

Autoimmune disorders are a highly heterogeneous class of conditions. Even within a single clinically diagnosed condition, such as SLE, the underlying causes and manifestations are highly variable. To better treat these disorders and transition toward a precision medicine framework, the last decade of research has used in-depth genetic and genomic studies to better resolve patient heterogeneity and identify the autoantigenome. The progress described in this review represents a substantial leap forward in both our understanding of these complex diseases and their potential treatments.

Although these are not the focus of the current review, there are several additional considerations that are important to note. For complex autoimmune diseases, environment, the interaction of genetics and environment as well as dietary, and lifestyle factors play an important role in affecting disease pathogenesis, progression, and treatment response. For instance, a recent study has shown that oral mucosal breaks trigger anti-citrullinated bacterial and human protein antibody responses in RA demonstrating the role of pathogens and environment in the disease (Brewer et al., Reference Brewer, Lanz, Hale, Sepich-Poore, Martino, Swafford, Carroll, Kongpachith, Blum, Elliott, Blachere, Parveen, Fak, Yao, Troyanskaya, Frank, Bloom, Jahanbani, Gomez, Iyer, Ramadoss, Sharpe, Chandrasekaran, Kelmenson, Wang, Wong, Torres, Wiesen, Graves, Deane, Holers, Knight, Darnell, Robinson and Orange2023). Studies focusing on molecular pathological epidemiology research, which can investigate those factors in relation to molecular pathologies and clinical outcomes have been explored for other conditions such as cancer (Hamada et al., Reference Hamada, Keum, Nishihara and Ogino2017; Hughes et al., Reference Hughes, Simons, van den Brandt, van Engeland and Weijenberg2017; Ogino et al., Reference Raffin, Vo and Bluestone2018). It is also important to note that the majority of existing studies in molecular profiling, genomics and genetics of autoimmune diseases have been carried out in patients of European background. If precision medicine is truly the goal, there is a need to explore social determinants of health in the context of disease progression and treatment response in diverse populations. More extensive studies are needed to explore the combination and interaction of the molecular, clinical, social, and environmental factors in diverse patient populations to achieve precision medicine for autoimmunity.

From using genetics to identify new gene targets, to using single-cell genomics to identify cellular and molecular subsets of disease, to computational approaches that aim to merge all this together and repurpose medicine in a targeted fashion, precision medicine in autoimmunity is an endeavor that will continue to yield enormous insights and lead to better – and more importantly – error-free therapeutics.