Roads and detours for CAR T cell therapy in autoimmune diseases

Abstract

Chimeric antigen receptor (CAR) T cell therapy has been highly effective in eradicating malignant B cells in cancer, and this success has prompted an extension of the approach to areas beyond oncology. In pioneering studies, CAR T cells targeting the B cell marker CD19 demonstrated robust efficacy as treatment for the autoimmune disease systemic lupus erythematosus. Patients who received anti-CD19 CAR T cells experienced remission of most or all clinical manifestations and discontinued prior medications. These results have spurred intense interest in extending these observations to larger patient cohorts and other autoimmune conditions. More nuanced strategies for use of CARs in autoimmunity have also been developed. Here, we offer insight into the role of B cells in the pathophysiology of autoimmunity and present an overview of preclinical studies and clinical trials that use engineered cell therapy for autoimmune disorders. In discussing the prospects and challenges of this emerging field, a view emerges in which the promise of clinical efficacy invites careful consideration of potential pitfalls.

Introduction

Chimeric antigen receptors (CARs) are engineered cell surface proteins used in immunotherapy. At the vanguard of synthetic medicine, CARs have garnered attention and fuelled excitement, as they instruct a patient’s own cells to aid in achieving therapeutic outcomes. In the years since the first patient with leukaemia received CAR T cells, the procedure has shown efficacy in thousands of patients with lymphoid malignancies and increasingly contributes to the standard of care. As a form of adoptive immunotherapy, CARs typically consist of the single-chain variable fragment (scFv) of an antibody, a transmembrane domain, intracellular signalling domains of the T cell receptor (TCR) CD3ζ chain and the functional domains of at least one co-stimulatory receptor (Fig. 1). When constitutively expressed on the surface of a T cell through viral transduction with engineered viral vectors, CARs enable T cells to recognize their target cells based on the antigen specificity of the scFv extracellular domain, whereas the cytoplasmic signalling and co-stimulatory domains activate the CAR T cells and initiate cell division^1,2. CAR T cell therapy targeting the B cell antigens CD19 or B cell maturation antigen (BCMA) has proven to be highly effective in eradicating malignant B cells in cancer^3,4. Six CAR T cell therapies have been approved for clinical use by the US Food and Drug Administration (FDA) — four targeting CD19 for B cell lineage leukaemia and/or lymphoma, and two targeting BCMA for multiple myeloma^4–6. In addition, CAR T cell therapies targeting CD22, another B cell antigen, are being evaluated in late-phase clinical studies⁷ (Box 1).

Fig. 1 |. Composite nature of CARs.

A prototypical chimeric antigen receptor (CAR) (centre panel) consists of an extracellular single-chain variable fragment (scFv) derived from an antibody (right panel), which gives the CAR its specificity to target an antigen; a transmembrane domain that anchors the CAR in the plasma membrane; and two signalling domains that activate the CAR T cell and initiate proliferation. The signalling domains are typically an intracellular signalling domain from the CD3ζ subunit of the T cell receptor (TCR) and a domain from the CD28 co-stimulatory receptor (left panel). APC, antigen-presenting cell; MHC, major histocompatibility complex.

Box 1 |. Status of CAR T cell therapies in oncology.

More than 75% of paediatric patients with relapsed and/or refractory B cell acute lymphoblastic leukaemia (B-ALL) achieved remission after treatment with anti-CD19 chimeric antigen receptor (CAR) T cells, with approximately 40–50% of those patients maintaining long-term remission^{9,12,13,16,19,204}. Similarly, clinical trials evaluating CAR T cell therapy in adults with relapsed/refractory B-ALL have seen the greatest success, followed by trials in relapsed/refractory lymphoma and multiple myeloma.

Anti-CD19 CAR T cell therapies have been evaluated in pivotal trials for adult patients with non-Hodgkin’s, diffuse large B cell or mantle cell lymphoma. Approximately 40–60% of these patients achieved a complete response post infusion. Additionally, responses were durable in a subset of patients, particularly those who were younger and had lower tumour burden at the time of treatment^{8,10,14,17,20}.

Subsequent studies evaluating the use of anti-CD19 CAR T cells as a second-line treatment for lymphoma in adults have also shown promising results compared with the standard of care²⁰⁵.

Targeting the B cell maturation antigen (BCMA) with CAR T cells has been successful in a subset of adult patients with relapsed/refractory multiple myeloma, with complete response rates ranging from 33% to 82.5%^206,207.

CAR T cell therapy targeting the B cell antigen CD22 has shown efficacy in adult patients with relapsed/refractory lymphomas, including patients previously treated with CD19-directed CAR T cell therapies²⁰⁸.

The development of these successful CAR therapies for cancer has relied on the use of the patient’s own (autologous) T cells to construct the CAR T cells. Clinical trials in oncology have employed various vector constructs, CAR co-stimulatory domains, lymphodepleting chemotherapy regimens prior to CAR T cell infusion, and CAR T cell doses and schedules. Despite these differences, the reported efficacy and toxicity profiles in different lymphoid malignancy indications have been similar and side effects generally manageable^8–20. Now, robust data from oncology clinics provide the blueprint for extending the applications of CAR T cells to autoimmune conditions.

Autoimmune diseases (AIDs) comprise several dozen heterogeneous disorders, and generally reflect the contributions of genetic factors, the environment and stochastic events occurring during lymphocyte maturation. In most cases, autoimmune pathogenesis involves crucial participation of the B cell lineages, which, by producing autoantibodies, often contribute directly to disease manifestations. B cells also function as potent antigen-presenting cells (APCs) that engage T cells and entrench autoimmunity. Additionally, B cells participate more broadly in directing the autoimmune response by secreting cytokines. Therefore, it has long been suspected that B cell depletion should alleviate the major manifestations of AIDs. However, the B cell-depleting anti-CD20 monoclonal antibody rituximab has shown inconsistent effects in clinical trials and many patients with AID remain inadequately treated²¹.

Conceptually, CAR T cells offer two advantages over monoclonal antibody therapy: first, CAR T cells proliferate in the patient, whereas a biologic agent that is injected into the bloodstream degrades over time; and second, CAR T cells actively migrate into diverse tissues to deplete tissue-resident B cells, whereas an antibody distributes by diffusion from the blood to the tissues²². Therefore, the CAR T cell therapies targeting B cell markers that were developed for oncology indications represented a promising approach for achieving deeper responses in AIDs. Detailed proof-of-principle experiments in animal models demonstrated that anti-CD19 CAR T cells were indeed effective in eliminating most pathological manifestations of pemphigus vulgaris (PV)²³, a skin blistering disorder, as well as symptoms of systemic lupus erythematosus (SLE)²⁴, a connective tissue disease that often involves multiple organ systems. These experiments confirmed that the use of CAR T cells to treat AIDs was more efficacious than using monoclonal antibodies²⁵.

Shortly after the mouse model data were published, the first patients with SLE were treated with autologous CAR T cells²⁶. The treatment proved remarkably effective at reversing disease progression and restoring organ functions. Ongoing clinical trials are evaluating autologous CAR T cells directed against CD19, CD20, CD22 or BCMA as treatments for B cell-mediated autoimmune disorders (Table 1 and Fig. 2), and promising early clinical data have been generated in several disorders. Although the number of reported studies is small and they are recently undertaken, most patients have achieved a remission of autoimmune manifestations or an overall improvement in symptoms, allowing them to discontinue medications. Overall, the risk–benefit profile has been favourable across different indications.

Table 1 |.

Clinical trials of CAR T cells in autoimmune diseases

Clinical condition	NCT number	Target	Cell type	Phase	Sponsor/PI
Lupus nephritis (LN)	NCT06904729	CD19	T cell	III	Guangzhou Women and Children’s Medical Center
Myasthenia gravis (MG)	NCT06799247	BCMA	T cell	III	Cartesian Therapeutics
ANCA-associated vasculitis (AAV)	NCT06868290	CD19	T cell	II	Novartis Pharmaceuticals
Autoimmune disease (AID)	NCT06794008	CD19 and BCMA	T cell	II	Peking University People’s Hospital
Stiff person syndrome (SPS)	NCT06588491	CD19	T cell	II	Kyverna Therapeutics
LN	NCT06581198	CD19	T cell	II	Novartis Pharmaceuticals
Multiple sclerosis (MS)	NCT06384976	CD19	T cell	II	Kyverna Therapeutics
MG	NCT06193889	CD19	T cell	II	Kyverna Therapeutics
Systemic lupus erythematosus (SLE)	NCT06038474	BCMA	T cell	II	Cartesian Therapeutics
MG	NCT04146051	BCMA	T cell	II	Cartesian Therapeutics
MS	NCT07006805	CD19	T cell	I/II	Cabaletta Bio
LN|systemic sclerosis (SSc)|Sjögren’s syndrome (SS)	NCT06947473	CD19 and BCMA	T cell	I/II	Beijing GoBroad Hospital
LN|SSc|SS	NCT06947460	CD19 and BCMA	T cell	I/II	Beijing GoBroad Hospital
LN	NCT06935474	CD19	T cell	I/II	AbelZeta
SLE	NCT06897930	CD19 and BCMA	T cell	I/II	AstraZeneca
SLE|LN	NCT06839976	CD19	T cell	I/II	Children’s Hospital of Philadelphia
SLE|SSc|SS|AAV|IIM|APS	NCT06828042	CD19	T cell	I/II	Peking University Third Hospital
SLE|SSc|IIM|AAV|SS	NCT06821659	CD19	T cell	I/II	Peking University Third Hospital
MG	NCT06704269	CD19	T cell	I/II	Novartis Pharmaceuticals
AID	NCT06688799	CD19	T cell	I/II	Beijing GoBroad Hospital
SLE|SSc|AAV|IIM	NCT06685042	CD19	T cell	I/II	Fondazione Policlinico Universitario Agostino Gemelli IRCCS
MS	NCT06675864	CD19	T cell	I/II	Novartis Pharmaceuticals
MS	NCT06617793	CD19	T cell	I/II	Novartis Pharmaceuticals
AAV	NCT06590545	CD19	T cell	I/II	David Simon
SLE	NCT06585514	CD19	T cell	I/II	Beijing GoBroad Hospital
SLE	NCT06530849	CD19 and BCMA	T cell	I/II	Gracell Biotechnologies
Rheumatoid arthritis (RA)	NCT06475495	CD19	T cell	I/II	Charite University
AID	NCT06435897	CD19 and BCMA	T cell	I/II	Shenzhen Geno-Immune Medical Institute
SLE	NCT06428188	CD19 and BCMA	T cell	I/II	Essen Biotech
SSc	NCT06400303	CD19	T cell	I/II	Kyverna Therapeutics
MG	NCT06359041	CD19	T cell	I/II	Cabaletta Bio
Immune thrombocytopenia (ITP)	NCT06352281	BCMA	T cell	I/II	920th Hospital of Joint Logistics Support Force of People’s Liberation Army of China
SLE|LN|AAV|GPA|MPA|SSc|IIM|SS	NCT06350110	CD19 and BCMA	T cell	I/II	Essen Biotech
SLE|SSc|IIM	NCT06347718	CD19	T cell	I/II	University of Erlangen-Nürnberg
LN	NCT06342960	CD19	T cell	I/II	Kyverna Therapeutics
SSc	NCT06328777	CD19	T cell	I/II	Cabaletta Bio
SLE	NCT06189157	CD19	T cell	I/II	Miltenyi Biomedicine GmbH
IIM	NCT06154252	CD19	T cell	I/II	Cabaletta Bio
SLE|LN	NCT06153095	CD19 and CD20	T cell	I/II	Lyell Immunopharma
SLE|LN	NCT06121297	CD19	T cell	I/II	Cabaletta Bio
SLE	NCT06106906	CD19	γδ T cell	I/II	Wuhan Union Hospital
SLE	NCT06106893	CD19	γδ T cell	I/II	Wuhan Union Hospital
LN	NCT05938725	CD19	T cell	I/II	Kyverna Therapeutics
SLE|LN	NCT05798117	CD19	T cell	I/II	Novartis Pharmaceuticals
AID	NCT05459870	CD19	T cell	I/II	Shenzhen Geno-Immune Medical Institute
MS	NCT07008378	CD19 and CD20	T cell	I	Genentech
SLE	NCT06984341	CD19 and CD20	T cell	I	Genentech
SLE|IIM|SSc|AAV	NCT06980597	NA	γδ T cell	I	Beijing GoBroad Hospital
SLE|AHA|MG|SSc|AAV|IIM|IgG4-RD	NCT06978738	CD19 and BCMA	T cell	I	Changzhou No. 2 People’s Hospital
MG	NCT06958939	BCMA	T cell	I	Ting Chang
SLE|SSc|IIM|AAV	NCT06941129	CD19 and BCMA	T cell	I	Institute of Hematology & Blood Diseases Hospital
SLE	NCT06934447	CD70 and BCMA	T cell	I	Children’s Hospital of Zhejiang University School of Medicine
SLE|LN|SSc|IIM	NCT06925542	CD19	T cell	I	CRISPR Therapeutics
SLE|SSc|IIM	NCT06913608	CD19	T cell	I	Calibr (Scripps Research)
SLE	NCT06892145	CD19	T cell	I	Chongqing Precision Biotech
MS|NMOSD|MOGAD|MG	NCT06869278	CD19	T cell	I	Wuhan Union Hospital
SLE	NCT06852573	CD19	T cell	I	Beijing Immunochina Medical Science & Technology
ITP	NCT06826430	CD19	T cell	I	Juventas Cell Therapy
SLE|SSc	NCT06822881	NA	T cell	I	Beijing GoBroad Hospital
ITP	NCT06787989	CD19 and BCMA	T cell	I	iCell Gene Therapeutics
MG	NCT06759948	CD19 and BCMA	T cell	I	Chongbo Zhao
SLE|LN	NCT06752876	CD19	T cell	I	Caribou Biosciences
AHA	NCT06733610	CD19 and BCMA	T cell	I	Institute of Hematology & Blood Diseases Hospital
SLE	NCT06710717	CD19	T cell	I	National University of Malaysia
LN|SLE|SSc	NCT06708845	CD19 and CD20	T cell	I	Miltenyi Biomedicine
SLE	NCT06694298	BCMA	T cell	I	CSPC ZhongQi Pharmaceutical Technology
SLE	NCT06691152	CD19	T cell	I	The Children’s Hospital of Zhejiang University School of Medicine
MG	NCT06688435	CD19	T cell	I	CSPC ZhongQi Pharmaceutical Technology
AID	NCT06680037	CD19	T cell	I	TG Therapeutics
AID	NCT06633042	BCMA and CD19	T cell	I	Bioray Laboratories
MG	NCT06626919	BCMA	T cell	I	Arcellx
AID	NCT06567080	CD19 and CD20	T cell	I	RenJi Hospital
LN|SLE|SSc	NCT06544330	CD19	T cell	I	Synthekine
SLE	NCT06465147	CD19	T cell	I	Seattle Children’s Hospital
MS	NCT06451159	CD19	T cell	I	Bruce Cree
SLE	NCT06429800	CD19	T cell	I	Atara Biotherapeutics
LN	NCT06375993	CD20	γδ T cell	I	Adicet
SLE	NCT06373991	CD19	T cell	I	EdiGene
MG	NCT06371040	CD19 and BCMA	T cell	I	Ting Chang
RA	NCT06201416	Citrullinated proteins	Treg cell	I	Sonoma Biotherapeutics
Hidradenitis suppurativa (HS)	NCT06361836	Citrullinated proteins	Treg cell	I	Sonoma Biotherapeutics
SLE	NCT06349343	CD19 and BCMA	T cell	I	Wuhan Union Hospital
SLE	NCT06340750	BAFF	γδ T cell	I	Luminary Therapeutics
SLE	NCT06340490	CD19	T cell	I	Guangdong Ruishun Biotech
SLE	NCT06333483	CD19	T cell	I	Autolus Limited
AAV|IIM|SSc|SLE	NCT06308978	CD19	T cell	I	Fate Therapeutics
Dermatomyositis	NCT06298019	CD19	T cell	I	Stanford University
SLE	NCT06297408	CD19	T cell	I	Shanghai Ming Ju Biotechnology
SLE|AAV|GPA|MPA	NCT06294236	CD19	T cell	I	Sana Biotechnology
LN|AAV|MN|IgG4-RD	NCT06285279	CD19 and BCMA	T cell	I	Nanjing University School of Medicine
SLE|IMNM|NMO|MS|MG	NCT06249438	CD20 and BCMA	T cell	I	RenJi Hospital
AHA	NCT06231368	CD19	T cell	I	Inst. Hematology & Blood D
SLE	NCT06222853	CD19	T cell	I	The Children’s Hospital of Zhejiang
MS	NCT06220201	CD19	T cell	I	Juno Therapeutics (BMS)
AHA	NCT06212154	CD19	T cell	I	Institute of Hematology & Blood Diseases Hospital
IIM|SSc|SLE|AAV	NCT06152172	CD19	T cell	I	David Porter
MS	NCT06138132	CD19	T cell	I	Stanford University
SLE|SS|SSc|dermatomyositis|AAV	NCT06056921	CD19	T cell	I	Chongqing Precision Biotech
SLE|IIM|SSc	NCT05869955	CD19	T cell	I	Juno Therapeutics (BMS)
SLE	NCT05846347	CD19 and BCMA	T cell	I	Zhejiang University
MG	NCT05828225	CD19	T cell	I	Zhejiang University
NMO	NCT05828212	CD19	T cell	I	Zhejiang University
SLE	NCT05765006	CD19	T cell	I	Shanghai Ming Ju Biotechnology
SLE	NCT05474885	CD19 and BCMA	T cell	I	iCell Gene Therapeutics
MG	NCT05451212	MuSK	T cell	I	Cabaletta Bio
Pemphigus vulgaris (PV)	NCT04422912	DSG3	T cell	I	Cabaletta Bio
SLE	NCT03030976	CD19	T cell	I	Shanghai GeneChem
AID	NCT06993493	CD19	T cell	NA	Beijing Boren Hospital
AID	NCT06983964	CD19	T cell	NA	Beijing Boren Hospital
MG	NCT06965309	CD19	T cell	NA	Shenzhen MagicRNA Biotechnology
SLE|LN	NCT06902844	BCMA	T cell	NA	Tongji Hospital
SLE	NCT06900764	NA	T cell	NA	Wuhan Union Hospital
AID	NCT06887985	CD19	T cell	NA	He Huang
SSc	NCT06871644	CD7	T cell	NA	PersonGen BioTherapeutics
SLE	NCT06801119	CD19	T cell	NA	Shenzhen MagicRNA Biotechnology
AID of the nervous system	NCT06797024	CD19	T cell	NA	Tongji Hospital
AID	NCT06661811	NA	T cell	NA	Beijing Boren Hospital
SLE	NCT06513429	CD19	T cell	NA	Peking University Third Hospital
AAV	NCT06508346	CD19	T cell	NA	The Children’s Hospital of Zhejiang University
AID	NCT06503224	CD19 and BCMA	T cell	NA	1st Affiliated Hospital of the University of Science and Technology
SLE|SS|SSc|IIM|AAV|APS	NCT06373081	CD19	T cell	NA	Shanghai Changzheng Hospital
SLE|IIM|SSc|IgG4-RD|SS	NCT06361745	CD19	γδ T cell	NA	PersonGen BioTherapeutics
SLE	NCT06310811	CD19	T cell	NA	Wuhan Union Hospital
LN|AAV	NCT06277427	BCMA	T cell	NA	Lingli Dong
SLE	NCT05988216	CD19	T cell	NA	Bioray Laboratories
SLE|SS|SSc|IIM|AAV|APS	NCT05859997	CD19	T cell	NA	Bioray Laboratories
IIM|AAV	NCT06986018	CD19	T cell	Early I	Peking University People’s Hospital
SLE	NCT06946485	CD70	T cell	Early I	The Affiliated Nanjing Drum Tower Hospital of Nanjing University Medical School
NMO|MG|MS|CIDP|AIE	NCT06939166	CD19	T cell	Early I	Tianjin Huanhu Hospital
MG	NCT06933563	CD19 and BCMA	T cell	Early I	Zhejiang University
AHA	NCT06920446	CD19 and BCMA	T cell	Early I	Zhejiang University
SLE	NCT06920433	CD19 and BCMA	T cell	Early I	Zhejiang University
SLE	NCT06886919	CD19	T cell	Early I	Beijing Immunochina Medical Science & Technology
AID	NCT06866080	CD19, CD20 and CD22	T cell	Early I	Nanjing Legend Biotech Co.
LN	NCT06785519	CD19 and BCMA	T cell	Early I	He Huang
SLE|SSc|IIM|NMOSD|MS|MG|AAV	NCT06775912	CD19 and BCMA	T cell	Early I	Nanjing Bioheng Biotech
AID	NCT06762119	NA	T cell	Early I	Zhejiang University
SLE	NCT06711146	CD19	T cell	Early I	Zhejiang University
IgAN|MN	NCT06690359	CD19	T cell	Early I	Beijing Immunochina Medical Science & Technology
LN	NCT06681337	CD19 and BCMA	T cell	Early I	Bioray Laboratories
AID	NCT06680388	CD19	T cell	Early I	Institute of Hematology & Blood Diseases Hospital
SLE	NCT06653556	CD19, CD20 and CD22	T cell	Early I	Wuhan Union Hospital
SLE|SSc|AAV|IIM|SS	NCT06549296	CD19	T cell	Early I	Nanjing Bioheng Biotech
AID	NCT06548620	CD19	T cell	Early I	Nanjing Bioheng Biotech
ITP	NCT06519565	BCMA	T cell	Early I	Wuhan Union Hospital
NMO|MG|MS|CIDP|AIE	NCT06485232	CD19 and BCMA	T cell	Early I	Xuanwu Hospital
SLE	NCT06420154	CD19	T cell	Early I	First Affiliated Hospital of Wenzhou
MG	NCT06419166	CD19 and BCMA	T cell	Early I	Zhejiang University
AID	NCT06417398	CD19	T cell	Early I	PersonGen BioTherapeutics
SLE	NCT06316791	CD19	T cell	Early I	Juventas Cell Therapy
SLE	NCT05930314	CD19	T cell	Early I	Peking Union Medical College Hospital
AID of the nervous system	NCT04561557	BCMA	T cell	Early I	Tongji Hospital
SLE|SSc|AAV|IIM|SS	NCT06548607	CD19	T cell	Early I	Nanjing Bioheng Biotech
LN|IgG4-RD	NCT06497387	BCMA	T cell	Early I	Tongji Hospital
IgG4-RD|LN	NCT06497361	CD19 and BCMA	T cell	Early I	Tongji Hospital
SLE|AAV|IIM	NCT06462144	CD19 and CD20	T cell	Early I	Nanjing University Medical School
AID	NCT06279923	CD19 and BAFF	T cell	Early I	Zhejiang University
SLE	NCT05858684	CD19 and BCMA	T cell	Early I	RenJi Hospital
POEMS|amyloidosis|AHA|vasculitis	NCT05263817	CD19 and BCMA	T cell	Early II	Zhejiang University
Crohn’s disease (CD)|UC| dermatomyositis|Still disease	NCT05239702	CD7	T cell	Early I	Zhejiang University
SSc	NCT05085444	CD19 and BCMA	T cell	Early I	Zhejiang University
SS	NCT05085431	CD19 and BCMA	T cell	Early I	Zhejiang University
SLE	NCT05030779	CD19 and BCMA	T cell	Early I	Zhejiang University
Immune nephritis	NCT05085418	CD19 and BCMA	T cell	Early I	Zhejiang University

AHA, autoimmune haemolytic anaemia; AIE, autoimmune encephalitis; ANCA, anti-neutrophil cytoplasmic antibody; APS, antiphospholipid syndrome; BAFF, B cell activation factor; BCMA, B cell maturation antigen; CAR, chimeric antigen receptor; CIDP, chronic inflammatory demyelinating polyradiculoneuropathy; DSG3, desmoglein 3; GPA, granulomatosis with polyangiitis; IgAN, Immunoglobulin A (IgA) nephropathy; IgG4-RD, IgG4-related disease; IIM, idiopathic inflammatory myopathy; IMNM, immune-mediated necrotizing myopathy; MN, membranous nephropathy; MOGAD, myelin oligodendrocyte glycoprotein antibody-associated disease; MPA, microscopic polyangiitis; MuSK, muscle-specific kinase; NA, not available; NMO, neuromyelitis optica; NMOSD, neuromyelitis optica spectrum disorder; PI, principal investigator; POEMS, POEMS syndrome; T_reg cell, regulatory T cell; UC, ulcerative colitis.

Fig. 2 |. Bubble plot of CAR T cell trials in autoimmune diseases, relating targets to indications.

A total of 164 chimeric antigen receptor (CAR) T cell clinical trials in autoimmune diseases (AIDs) were identified from ClinicalTrials. gov in May 2025. Trials have been grouped by the CAR target (y axis) and clinical indications (x axis). AAV, ANCA-associated vasculitis; AHA, autoimmune haemolytic anaemia; AIE, autoimmune encephalitis; ANCA, anti-neutrophil cytoplasmic antibody; APS, antiphospholipid syndrome; BAFF, B cell activation factor; BCMA, B cell maturation antigen; CD, Crohn’s disease; CIDP, chronic inflammatory demyelinating polyradiculoneuropathy; CNS, central nervous system; DSG3, desmoglein 3; GPA, granulomatosis with polyangiitis; HS, hidradenitis suppurativa; IgAN, Immunoglobulin A (IgA) nephropathy; IgG4-RD, IgG4-related disease; IIM, idiopathic inflammatory myopathy; IN, immune nephritis; ITP, immune thrombocytopenia; LN, lupus nephritis; MG, myasthenia gravis; MN, membranous nephropathy; MOGAD, myelin oligodendrocyte glycoprotein antibody-associated disease; MPA, microscopic polyangiitis; MS, multiple sclerosis; MuSK, muscle-specific kinase; NMOSD, neuromyelitis optica spectrum disorder; POEMS, POEMS syndrome; PV, pemphigus vulgaris; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SPS, stiff person syndrome; SS, Sjögren’s syndrome; SSc, systemic sclerosis; UC, ulcerative colitis.

The positive results for CAR T cell treatments in AIDs were achieved with milder side effects than those observed in lymphoid malignancies (Box 2). A comparison of potential anti-CD19 CAR T cell toxicities in patients with B cell malignancies versus those with SLE revealed a lower incidence of cytokine release syndrome (CRS), immune-effector cell-associated neurotoxicity syndrome, and haematotoxicity in the autoimmune group²⁷. This could be attributed to several possible factors, including fewer target cells, more robust condition of the harvested autologous T cells, and the patients’ youth and often better overall health. In patients with SLE, the CAR T cell therapy showed a deep depletion of CD19⁺ B cells that lasted approximately 3 months and was followed by reconstitution of B cell subsets.

Box 2 |. Caveats emerging from CAR T cell therapies for cancer.

Relapse has been observed following initial remission after chimeric antigen receptor (CAR) T cell therapy. The main relapse mechanisms are shared across multiple types of B cell malignancies. Recurrence of cancer can be associated with the emergence of variants that lead to loss of the B cell antigen^209–212, with CAR T cell expansion and persistence being only transient²¹³ and with lineage switching away from cells expressing the targeted antigen²⁰⁴. Potentially, antigen loss variants could be deleted by targeting CD22, a second antigen expressed on B cells, a treatment that is showing high initial response rates^214–216 (see Box 1). Promising results were similarly observed in studies using CARs directed against CD20 (ref. 217) or using CD19–CD20 bispecific CARs^218,219.

Despite their efficacy in inducing remission in heavily pretreated patients, CAR T cell therapies are associated with the risk of significant short-term morbidity and mortality. Due to immune system activation after CAR T cell infusion, patients can experience varying degrees of systemic inflammation, which can lead to clinical symptoms, including cytokine release syndrome (CRS), immune-effector cell-associated neurotoxicity syndrome and/or immune-effector cell-associated haemophagocytosis-like syndrome (IEC-HS)^220–222.

CRS is a systemic inflammatory response which typically occurs within 1–2 weeks post CAR T cell infusion, correlating with the timing of maximal CAR T cell expansion. The most common manifestations of CRS include persistent fever, potentially accompanied by hypotension, hypoxia and organ damage. Generally, CRS can be medically managed with supportive care and is treated with monoclonal antibodies targeting the interleukin-6 (IL-6) receptor (tocilizumab), or elevated cytokines, or with systemic immunosuppression^221,222.

Neurologic changes after administration of CAR T cells have also been reported²²³. Neurotoxicity resulting from CD19 CAR T cell infusion is thought to be related to CAR T cell proliferation and the subsequent release of pro-inflammatory cytokines, monocyte/macrophage activation, endothelial cell activation, disruption of the blood–brain barrier and, potentially, off-tumour, on-target toxicity for structurally related neurologic targets^224–228. The spectrum of observed clinical symptoms of toxicity is quite broad across individuals and can include transient language disturbances, alterations in consciousness, cognitive impairment, confusion and rare severe features including seizures, paralysis or fatal cerebral oedema^224–228.

More recently, IEC-HS has emerged as an entity distinct from CRS in the spectrum of immune dysregulation after CAR T cell therapy. Similar to haemophagocytic lymphohistiocytosis or macrophage activation syndrome, IEC-HS is often associated with cytopenias, multiorgan dysfunction and coagulopathy²²⁹.

Although the use of CAR T cells for autoimmunity is quite recent, accumulating evidence suggests that IEC-HS can also occur in this cohort of patients, albeit with differing severity and manifestations. The incidence and severity of CRS and immune-effector cell-associated neurotoxicity syndrome appear to be milder in autoimmune settings due to a lower overall immune activation level and the absence of tumour burden^{55,118,119,230}.

In this Review, we explain the features of AIDs that indicate their suitability for CAR T cell therapy. We provide an update on the remarkable progress of the clinical trials applying CAR T cells for autoimmune disorders, and argue for the continued focus on basic research and the importance of developing further applicable preclinical models. We end with an outlook on the hurdles and opportunities in this emerging field of engineered cell applications for AID therapy.

B cells in autoimmune pathophysiology

B cells originate from the bone marrow and undergo selection at various immune tolerance checkpoints during their development. The selection process begins with the expression of a unique B cell receptor (BCR) and culminates in the formation of affinity matured antibodies, which are secreted by terminally differentiated plasma cells^28–30. The formation of antibody-secreting plasma cells mostly occurs in germinal centres, where T cells and B cells interact. Concurrently, B cells undergo clonal expansion and multiple rounds of somatic hypermutation of the BCR, resulting in selection for high-affinity binding clones. However, B cells can also mature outside the germinal centres³¹, via an extrafollicular pathway that yields antibody-secreting cells with an important role in AIDs and severe infections^32–34.

Several B cell subsets have been observed at increased frequencies in AIDs, including plasmablasts, CD19 bright memory B cells, double-negative (DN2, CD27⁻IgD⁻) cells, age-associated B cells (ABCs), CD5⁺ B cells and atypical memory B cells^35,36. In the early phases of disease, DN2 B cells and ABCs differentiate into antibody-forming cells that amplify autoimmune responses and inflammation, given their ability to respond to autoantigens in conjunction with innate stimuli such as ligands for the Toll-like receptors TLR7 and TLR9 (refs. 33,37). The activation and persistence of memory B cells and plasma cells are considered critical for the chronic phase of AIDs^38,39.

Comprehensive analyses have characterized the cell surface markers that distinguish B cell subsets in healthy individuals (Table 2). In dysregulated states, these baseline distributions are altered, and expanded B cell subsets have been observed as common features of many AIDs (Table 3). These population biases can affect the full spectrum of B cell development, from early precursors to antibody-secreting plasma cells, which may represent opportunities for more targeted therapies. An ideal scenario would be to selectively eliminate the ‘harmful’ B cell populations while preserving the B cells that are essential for healthy immune function. Yet this goal is not always straightforward. Autoreactive B cell clones often occur across multiple subsets or reside in different tissue compartments, making it difficult to fully eliminate them with a single targeting approach^40,41.

Table 2 |.

Markers and characteristics of B cell subsets

B cell subset	Core markers	Key markers	Characteristics
Pro-B cell	CD19+, CD20−, CD10+, CD38+, CD21−, IgM−, IgD−	CD34+, TdT+	Early precursor in bone marrow; no immunoglobulin rearrangement
Pre-B cell	CD19+, CD20+, CD10+, CD38+, CD21−, IgM−, IgD−	CD34−, cytoplasmic μ (cμ+)	Pre-BCR expressed (μ-chain, surrogate light chain)
Immature B cell	CD19+, CD20+, CD10+, CD38+, CD21−, IgM+, IgD−	Surface IgM+, IgD−	First surface IgM; negative selection begins
Transitional B cell	CD19+, CD20+, CD10+, CD38++, CD21+/−, IgM+, IgD+/−	CD27−, CD10+	Early bone marrow emigres entering periphery
Naive mature B cell	CD19+, CD20+, CD10−, CD38−/low, CD21+, IgM+/−, IgD++	IgD++, CD27−	Mature, circulating naive B cell
Activated B cell	CD19+, CD20+, CD10−, CD38−/low, CD21−, IgM+/−, IgD+/−	CD69+, CD86+, MHC-II++, CD71+	Upregulated MHC class II and co-stimulatory molecules
Germinal centre B cell	CD19+, CD20+, CD10+, CD38+, CD21−, IgM−, IgD−	BCL6+, CXCR5+, Ki67+, AID	Somatic hypermutation and class switch recombination
IgM+ memory B cell	CD19+, CD20+, CD10−, CD38−/low, CD21+, IgM+, IgD+/−	CD21+, CD27+ (subset CD27−)	T cell-dependent and T cell-independent responses
Plasmablast B cell	CD19+ (reduced), CD38+, CD27+, CD44+, CD24low	CD44+, CXCR4	Proliferate rapidly and appear usually transiently in blood

AID, autoimmune disease; BCR, B cell receptor; MHC, major histocompatibility complex.

Table 3 |.

Relationship between B cell subset distribution and disease manifestations of selected autoimmune disorders

Autoimmune disease	Increased B cell subsets	Clinical relevance	Refs.
Systemic lupus erythematosus (SLE)	CD27−IgD− (double-negative (DN2)) B cells, ABCs/atypical B cells, plasmablasts	Correlate with disease flares, autoantibody production and poor prognosis	33,188
Rheumatoid arthritis (RA)	Memory B cells in synovium, CD27+ switched memory B cells, plasma cells, ABCs	Drive joint inflammation and autoantibody (for example, RF, anti-CCP) production	189–191
Multiple sclerosis (MS)	CD19+CD27+ memory B cells, B cells in CSF, plasmablasts during relapses, ABCs	Contribute to CNS inflammation and oligoclonal band formation	192,193
Pemphigus vulgaris (PV)	Desmoglein-specific memory B cells, plasmablasts, autoreactive B cells, ABCs	Essential for production of anti-desmoglein IgG autoantibodies	194,195
Type 1 diabetes	CD20+CD27+ memory B cells, islet-antigen-specific B cells, B naive D (BND) B cells	Support autoantibody responses to islet antigens (for example, IAA, GADA, IA2, ZnT8)	196,197
Myasthenia gravis (MG)	CD19+CD27+ memory B cells, plasmablasts, anti-AChR autoreactive clones, ↑ ABCs	Facilitate production of anti-AChR antibodies and disease exacerbations	198,199

ABC, age-associated B cell; AChR, acetylcholine receptor; CCP, cyclic citrullinated peptide; CNS, central nervous system; CSF, cerebrospinal fluid; RF, rheumatoid factor.

Upon activation, self-reactive B cells form large clones, switch isotypes and differentiate into antibody-secreting plasmablasts or plasma cells that secrete autoantibodies. Autoantibodies are pathologic hallmarks of specific diseases, such as SLE, and can precede the development of clinical signs of AIDs for several years^42,43. Some autoantibodies can have pathogenic effects by binding directly to tissues: notably, anti-desmoglein 3 (anti-Dsg3) antibodies promote skin blistering in PV⁴⁴ and anti-GPIIb/IIIa antibodies promote platelet opsonization in immune thrombocytopenic purpura^45,46. Other autoantibodies cause damage by forming immune complexes that deposit in the joints, as seen in rheumatoid arthritis (RA)⁴⁷, or along the basement membrane of the kidney glomeruli, as occurs in Goodpasture syndrome⁴⁸. Such antibody deposits promote complement activation and inflammation, and disease severity can be further enhanced by differential binding to variants of Fc receptors⁴⁹. Autoantibodies can also antagonize cellular functions; for example, the anti-acetylcholine receptor (anti-AChR) antibodies in myasthenia gravis (MG) impede the normal transmission of nerve impulses across neuromuscular junctions⁵⁰.

Self-reactive B cells can also capture self-antigens and act as APCs, stimulating self-reactive T cell responses. Unlike other APCs, B cells recognize specific, often dilute, antigens through their BCR and then present such antigens to T cells. The antigen-specific manner of the T cell–B cell interaction can amplify pathogenic T cell responses. In a mouse model of SLE, the essential role for the BCR was demonstrated by disabling the secretion of antibodies. Despite the absence of secreted antibody, the mice developed lupus manifestations⁵¹. Similarly, the disruption of B cell antigen presentation ameliorated diabetes in the non-obese diabetic (NOD) mouse model⁵². Additionally, B cells secrete pro-inflammatory cytokines such as interleukin-6 (IL-6) and GM-CSF, which, in diseases such as multiple sclerosis (MS), can influence the pathogenic effector functions of other innate and adaptive immune cells and contribute to overall pathogenesis^53,54.

During flares in several AIDs, short-lived plasmablasts and activated B cells boost autoantibody production and fuel inflammation. Anti-CD19 therapies might therefore offer advantages over anti-CD20 biologic agents because plasmablasts and memory B cells maintain expression of CD19 whereas CD20 expression decreases. CD19-targeted strategies thus eliminate a broader range of B cells and antibody-producing cells, including cells that escape CD20-based depletion (Table 2). Accordingly, CD19-directed CAR T cell therapy induced lasting remissions in patients with SLE who were resistant to CD20-based depletion therapy⁵⁵. The effectiveness of this approach argues that CD19⁺ plasmablasts, rather than CD19⁻ cells, are the key source of pathogenic autoantibodies in SLE, and presumably other connective tissue disorders.

Importantly, AIDs are not all driven by the same mechanisms. Although most are dominated by self-reactive B cells and autoantibodies, others rely more heavily on autoreactive T cells, or on the interaction between B cells and T cells. In T cell-dependent autoimmune disorders, depleting a wider range of B cell populations, including those involved in presenting antigens or producing inflammatory cytokines, might provide additional benefits beyond lowering autoantibody levels. Because different B cell markers vary in their expression and biological roles, newer CAR T cell therapies that target more defined B cell subsets might help overcome resistance mechanisms and provide more effective control of complex AIDs.

Contributions of B cells to specific autoimmune diseases

Autoreactive B cells are likely to contribute to disease progression in most AIDs and can serve as suitable diagnostic markers (Table 3). In this section, we highlight some of the unique roles of B cells in selected AIDs and discuss the disease characteristics that are relevant for B cell-targeting therapies, including CAR T cells.

Systemic lupus erythematosus

SLE is a B cell-mediated AID in which autoantibody formation against nucleic acids and nucleic acid binding proteins is a defining feature and important driver of pathophysiology⁵⁶. In patients with SLE with high disease activity, DN2 B cells are predominantly autoreactive and exhibit heightened expression of nucleic acid sensors, such as TLR7, and the B cell activation factor (BAFF) receptor, which promotes their selection and expansion⁵⁷. In response to elevated levels of type I interferons, BAFF and IL-21, the maturation of DN2 cells tends to follow an extrafollicular pathway. Such extrafollicular maturation leads to the development of short-lived antibody-producing plasmablasts that are typical for SLE⁵⁸. These plasmablasts display high levels of the CD19 antigen and low expression of CD20, which might account for the remarkable efficacy of the anti-CD19 CAR T cells⁵⁵. The role of long-lived CD19^lowBCMA⁺ plasma cells, which derive from typical germinal centre-based follicular B cell maturation pathways, is also well described and might help guide the development of anti-BCMA CAR therapies for certain patient subgroups⁵⁹.

Rheumatoid arthritis

RA is an AID that typically affects the joints but also impacts the visceral organs. B cells are deeply implicated in its pathophysiology, as autoantibodies including rheumatoid factor (RF) and anti-citrullinated peptide antibodies are directly linked to disease progression and are diagnostic for the disease⁶⁰. Within the most aggressively affected joints, B cells, plasmablasts and plasma cells are often detected and can form aggregates resembling ectopic lymphoid structures⁶¹. Therefore, the depletion of B cells with rituximab is an FDA-approved and effective therapeutic strategy in RA. However, the depletion of B cells in tissues with rituximab is limited, and bispecific T cell engagers that attract T cells via CD3 binding and guide them towards CD19 on B cells have demonstrated enhanced efficacy⁶². Furthermore, blocking key cytokines released by B cells, such as IL-6, or disrupting T cell–B cell interactions using the CTLA4 fusion protein abatacept is efficacious in many patients with RA⁶³.

Systemic sclerosis

Systemic sclerosis (SSc) is characterized by immune dysregulation, fibrosis and vasculopathy. B cells play a central role in its pathogenesis, as B cell activation is driven by an imbalance in regulatory pathways, resulting in the upregulation of CD19 and elevated B cell survival signals, such as BAFF⁶⁴. Disease progression involves multiple mechanisms including direct B cell-mediated fibroblast activation⁶⁵, the production of pro-inflammatory and pro-fibrotic cytokines, such as IL-6, and the excessive production of extracellular matrix that stimulates fibrosis^65,66. Disease-associated autoantibodies — including anti-topoisomerase I and anti-RNA polymerase III antibodies — not only correlate with disease activity and severity but can also amplify endothelial cell activation and damage, thereby exacerbating vasculopathy, a hallmark of the disease^64,67. Therapeutically, B cell depletion has shown promise in modifying the course of SSc, although the effectiveness of anti-CD20 monoclonal antibodies varies among patients⁶⁸. In recent studies, anti-CD19 CAR T cells have shown remarkable efficacy in treating refractory sclerosis^69,70.

Pemphigus vulgaris

B cells are central to the pathogenesis of pemphigus and produce autoantibodies targeting desmosomal cadherins termed desmogleins⁷¹. These autoantibodies directly cause acantholysis and blistering, as demonstrated in a pemphigus in vivo model⁷². Concordantly, adoptive transfer of pemphigus IgG, either as purified IgG from sera or as recombinant monoclonal IgG, is both necessary and sufficient to induce acantholysis and blister formation in experimental animals or organ cultures of human skin^72,73. In notable proof-of-principle experiments, mice were treated with CAR T cells constructed using extracellular domains of the desmoglein protein Dsg3, which acted as the autoantigenic ‘bait’ for the self-reactive B cells. These studies stimulated ongoing clinical trials for PV that use desmoglein fragments as the extracellular portion of the chimeric receptor²³.

Myasthenia gravis

MG is a chronic, autoimmune neuromuscular disease characterized by fluctuating skeletal muscle weakness and fatigue. In advanced, severe cases, it can lead to paralysis and an inability to swallow or breathe. This disorder is most often associated with autoantibodies against AChRs⁷⁴ or a muscle-specific kinase (MuSK)⁷⁵. There are several notable clinical trials that are exploring the use of CAR T cells for MG⁷⁶ (Table 1).

ANCA-associated vasculitis

Anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a rare, life-threatening AID characterized by inflammation and damage to small blood vessels throughout the body. AAV can affect various organs, most commonly the kidneys, lungs, skin, nerves and upper respiratory tract⁷⁷. The defining autoantibodies in AAV primarily target the neutrophil cytoplasmic antigens myeloperoxidase and proteinase 3 (ref. 78). Anti-CD19 CAR T cells are among a growing number of experimental therapies for this disorder, with several in clinical trials^79–81 (Table 1).

Type 1 diabetes mellitus

Historically, type 1 diabetes mellitus has been considered a T cell-mediated disease, due to cytotoxic T cell infiltration in the pancreas and the destruction of β-cells, which produce insulin⁸². The result, insulin deficiency, is associated with numerous lifelong comorbidities, including cardiovascular disease, retinopathy and diabetic kidney disease. Despite the direct role of cytotoxic T cells in diabetes pathology, increased attention has focused on the role of B cells, which act as potent APCs and as a source of autoantibodies and pro-inflammatory cytokines⁸³. These observations have motivated the use of rituximab in diabetes, which showed clinical benefit and preserved islet functions in treated individuals⁸⁴. Preclinical studies are exploring strategies for the use of CARs in mouse models of type 1 diabetes⁸⁵.

Conventional therapeutic approaches for autoimmune disease

Traditional therapeutic approaches for AIDs are aimed at reducing the activity of immune cells but typically fail to restore self-tolerance. Therapies include glucocorticoids and disease-modifying anti-rheumatic drugs (DMARDs), such as monoclonal antibodies against pro-inflammatory cytokines or small molecule inhibitors of intracellular signalling pathways. With the increased availability of biologics, monoclonal antibodies that directly target B cells have been tested for different AIDs but have shown variable efficacy.

Rituximab has demonstrated remarkable efficacy in treating RA, immune thrombocytopenia (ITP), AAV and MS^21,86–88. This has led to its FDA approval for RA, granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA) and PV^89,90. The primary effect of rituximab is the rapid depletion of CD20⁺ B cells, including memory B cells, that results in long-lasting clinical and serological remission⁹¹. B cells bound by rituximab are killed via antibody-dependent cellular cytotoxicity, macrophage clearance or complement activation⁹².

Given the clinical efficacy of rituximab, other monoclonal antibodies have been developed to deplete B cells. These contain fully human variable region gene sequences and/or glycoengineering modifications that lead to increased antibody-dependent cellular cytotoxicity of B cells. The new generation of monoclonal antibodies includes ofatumumab, ocrelizumab and obinutuzumab (antibodies against CD20), ianalumab (an anti-BAFF-R antibody) as well as blinatumomab (a CD19-targeted T cell engager) and mosunetuzumab (a CD20-targeted T cell engager)⁹³. These biologics are under investigation in clinical trials for various AIDs after having shown utility for the treatment of B cell malignancies. In addition to B cell-specific targeting antibodies, the plasma cell targeted anti-CD38 antibody daratumumab has shown strong efficacy in multiple AIDs⁹⁴.

However, these antibody-mediated approaches to B cell depletion have several limitations. For example, although generally well tolerated, rituximab can be associated with infusion reactions, B cell depletion-related infections^95,96, hypogammaglobulinaemia (particularly in patients with long-term B cell depletion)⁹⁷ and, in rare instances, progressive multifocal leukoencephalopathy⁹⁸. Additionally, some patients may require periodic retreatment due to the variable duration of rituximab’s effects.

Despite the significant advances in the treatment of autoimmune conditions, the complex pathophysiology of most disorders, their often-progressive course and the large inter-patient variability in disease manifestations result in variable responses to treatment. Alternative treatment options are thus needed in cases of refractory disease⁹⁹. Often, effective treatments still leave most patients with burdensome residual symptoms and tissue damage, which significantly interfere with the quality of life.

The therapeutic management of autoimmune conditions, which are chronic, syndromic and frequently associated with multiple comorbidities, often requires the concomitant administration of corticosteroids leading to long-term tissue damage¹⁰⁰. Thus, there is still a large need to develop safer treatments. Finding broadly effective therapeutic options across the spectrum of different autoimmune disorders and disease manifestations is complicated because each disorder has specific challenges related to its biology, severity and pathophysiology. This is exemplified by difficulties in fully characterizing the disease mechanism driving SLE¹⁰¹. Because of the heterogeneous disease presentation and the inadequate therapeutic efficacy of rituximab²¹, it has remained a matter of debate whether B cells are the main cell type driving lupus pathogenesis.

Preclinical models for CAR T cells in autoimmune diseases

The incomplete consensus about the pathogenetic induction and progression in SLE required initial testing of engineered cell therapies to be carried out in mouse models of lupus. Fortunately, biologically analogous, spontaneous models of SLE are available^102,103 and anti-murine CD19 CAR T cells had been developed and shown to be effective against murine B cell leukaemia in vivo¹⁰⁴. This allowed the efficacy of anti-CD19 CAR T cells as therapy for lupus to be examined²⁴, and, subsequently, several preclinical models of AIDs have yielded important insights into the potential impact of cell therapies (Table 4).

Table 4 |.

Mouse models of autoimmune disorders, in which CAR T cell therapies have been explored

Disease	Model	CAR	Target antigen	Outcome	Ref.
Type 1 diabetes mellitus	WT and immunodeficient NOD mice	Anti-InsB-g7 CAR Treg cells	Insulin B peptide-MHC complex	Prevented development of diabetes in NOD mice	85
Inflammatory bowel disease	TNP-specific CR transgenic mice	Anti-TNP CAR Treg cells	TNP	Improved survival when administered before or after disease induction	200
Multiple sclerosis (MS)	Mice immunized with rhMOG	Anti-MOG CAR Treg cells	MOG	Resolved disease and prevented relapse	201
Myasthenia gravis (MG)	EAMG mice	Anti-MuSK CAR T cells and anti-CD19 CAR T cells	MuSK or CD19	Disease resolution in both groups but less B cell depletion in MuSK-targeted group	202
NMDAR encephalitis	Passive transfer NMDAR encephalitis mice	Anti-NMDAR CAR T cells	NMDAR	Autoantibody production and CNS penetrance reduced	203
Systemic lupus erythematosus (SLE)	(NZB×NZW) F1 and MRLfas/fas	Anti-CD19 CAR T cells	CD19	Reversed disease and improved survival significantly	24
Systemic sclerosis (SSc)	Fra-2 transgenic mouse model	Anti-CD19 CAR T cells	CD19	Worsened pulmonary disease and failed to show efficacy	105
Pemphigus vulgaris (PV)	Modified active immunized PV mice	Anti-DSG3 CAR T cells	DSG3	Efficacy in targeting human anti-DSG3 B cells	71

CAR, chimeric antigen receptor; CNS, central nervous system; DSG3, desmoglein 3; EAMG, experimental autoimmune myasthenia gravis; MHC, major histocompatibility complex; MOG, myelin oligodendrocyte glycoprotein; MuSK, muscle-specific kinase; NMDAR, N-methyl-D-aspartate receptor; NOD, non-obese diabetic; rhMOG, recombinant human myelin oligodendrocyte glycoprotein; TNP, 2,4,6-trinitrophenol; T_reg cell, regulatory T cell; WT, wild type.

Mouse models have also been valuable for the development of more specific strategies to deplete autoantibody-producing B cells. In studies of PV, a chimeric autoantigen-expressing T cell was designed to specifically eliminate Dsg3-reactive B cells (Fig. 3). This chimeric autoantibody receptor (CAAR) T cell expresses the epitopes of Dsg3 that specifically bind autoreactive B cells, whereas the intracellular portion of the receptor contains T cell stimulatory domains from CD137 and CD3ζ to induce cytotoxic responses towards Dsg3-reactive B cells²³. Dsg3-targeted CAAR T cell therapy is now being investigated in clinical trials for PV⁷¹ (Table 1).

Fig. 3 |. Alternative CAR T cell approaches.

a, In the prototypical chimeric antigen receptor (CAR) T cell approach, autologous T cells from a patient are transduced with DNA vector-based CARs that target B cell antigens, such as CD19, CD20, CD22 and B cell maturation antigen (BCMA), and infused back into the patient. b, Alternatively, in chimeric autoantibody receptor (CAAR) T cell approaches, autologous T cells are transduced with CARs that display autoantigen epitopes (such as desmoglein 3 (Dsg3) ectodomains) rather than single-chain variable fragment (scFv), which specifically target pathogenic B cells while sparing beneficial cells. c, Allogeneic T cells from healthy donors can be utilized for CAR transduction following knockout of genes encoding the major histocompatibility complex (MHC) and T cell receptor (TCR). These engineered CAR T cells are subsequently infused into patients. d, In the in vivo approach, mRNA constructs encoding CARs can be delivered to patients using lipid nanoparticles (LNPs), so that transduction of T cells occurs in vivo to form the CAR T cell. BCR, B cell receptor.

Experiments in mouse models also help expose potential problems that might arise with cell therapies. In a transgenic mouse model of SSc, CAR T cells effectively depleted B cells but worsened disease outcomes, including increased lung fibrosis, pulmonary hypertension and mortality¹⁰⁵. This deterioration was linked to the accumulation and activation of CAR T cells in tissues with a high T cell burden, leading to excessive cytokine production and systemic inflammation. There could be two reasons for these observations: first, the Fra-2 overexpressing mouse strain used in this system is an inexact model of the human disorder; and second, the CAR T cells were administered without prior lymphodepletion, increasing the potential pro-inflammatory effect of the treatments. Importantly, anti-CD19 CAR T cell therapy for SSc has shown efficacy in human trials⁷⁰, yet caveats inferred from animal models should be considered for potential risks in a subset of patients with AID.

Animal models also present opportunities to test the effects of targeting specific B cell subsets. ABCs, mentioned above, are characterized by the expression of the ZEB2 transcription factor across different species^106,107. Transcriptional profiling of ABC-like cells and the related DN2 subset revealed the surface immunoregulatory protein Fc receptor-like 5 (FCRL5) as a potential candidate antigen^57,108,109. An alternative potential target is SLAM7, which is similarly upregulated in ABCs and is already an FDA-approved immunotherapy target for multiple myeloma¹¹⁰. The shared distribution and regulation of these two receptors in humans and mice lends credence to their promise as well as to the utility of preclinical testing strategies.

However, there are potential pitfalls in using animal models, which include interspecies differences in disease features, pathology, genetics and regulatory mechanisms. For example, despite their conserved distribution within the B cell compartment and in ABC-like populations present in autoimmune states, FCRL5 homologues exhibit marked sequence and structural differences between humans and mice¹¹¹. Nevertheless, animal models present opportunities for fine-tuning therapeutic strategies before advancing into human trials.

Clinical testing of CAR T cells for autoimmune disease

Rapidly following the preclinical tests, case studies in patients demonstrated the transformative potential of CAR T cell therapy¹¹². Individual case studies and ongoing clinical trials are investigating CAR T cell therapies across a range of AIDs, including SLE^{26,55,113–115}, MS^116,117, lupus nephritis (LN)¹¹⁸, SSc^55,119, Sjögren’s syndrome (SS)¹²⁰, antisynthetase syndrome^121,122, myasthenia gravis (MA)^123,124 and idiopathic inflammatory myopathies^70,125,126 (Fig. 2). As of May 2025, 164 clinical trials that involve using CAR technology for autoimmune indications were listed in the ClinicalTrials.gov database (Table 1 and Fig. 2).

More than a quarter of these trials test efficacy in patients with SLE or LN. Nearly half of the trials applied CAR T cells that target CD19, with additional CAR targets including CD20, BCMA or BAFF. In addition, dual-targeting CAR T cells that use bispecific approaches to bind CD19–BCMA, CD19–CD20 or CD20–BCMA are being tested in the clinic. Other CAR T cell trials are targeting CD7 on autoreactive T cells for conditions including Crohn’s disease (CD), ulcerative colitis (UC), dermatomyositis and Still’s disease (Table 1).

The early consensus from clinical trials in AIDs is that CAR T cell therapies are generating promising preliminary data in SLE, LN, myositis and generalized MG, achieving clinical remission or symptom improvement. Across indications, post-CAR treatment analysis revealed that laboratory parameters normalized in treated patients. Mackensen et al. reported that patients with SLE whose disease went into remission did not need further immunosuppression and could stop therapy for more than a year¹¹⁴. This remission was attributed to a complete reset of the B cell repertoire driven by a complete depletion of B cells in tissues (documented by lymph node biopsy¹²⁷), followed by the repopulation of naive B cells (documented by flow cytometry and longitudinal BCR sequencing^55,128). Remarkably, a patient with severe, refractory SLE including LN who was treated with autologous CD19 CAR T cells is now without signs of autoimmunity or disease relapse for more than 3 years¹²⁹. In a progress update, 39 patients treated with CAR T cells have shown an initial response to therapy and only one relapse¹³⁰. The anti-CD19 CAR T cell approach has also shown efficacy in adolescent individuals experiencing rapidly progressive, severe and refractory SLE^118,131. Patient follow-up revealed a new type of toxicity associated with CD19-targeted CAR T cells, which was termed local immune-effector cell-associated toxicity syndrome. This possible side effect was generally mild and responded to glucocorticoid treatment¹³⁰.

More broadly, the clinical experience with CAR T therapies for AIDs has thus far shown them to be generally safe, with primarily low-grade adverse events. However, the long-term outcomes (including the potential for negative effects of the conditioning regimen) remain largely unknown.

CAR design informed by T cell signalling

Increased efficacy coupled with lower side effects is the major goal in the evolving development of CAR T cells for the treatment of autoimmunity. A nuanced understanding of T cell signalling is required to design efficacious CAR T cells. TCRs are multisubunit signalling complexes poised to transduce intracellular activation upon cognate antigen engagement^132,133. Accordingly, CAR signalling is based on principles of T cell activation and effector functions, yet the chimeric nature of CARs uncovers new paradigms of multipartite function.

CAR T cell efficacy is impacted by the overall CAR surface density, as well as the distinct structural elements of the CAR including the antigen-binding domain, the linker segment between the VH and VL, the length and flexibility of the hinge connecting the scFv to the transmembrane domain, the transmembrane domain, the co-stimulatory domains and the TCR CD3ζ signalling tail¹³⁴. Each of these has shown potential to alter the efficacy of CAR T cell therapies in AIDs.

The initial persistence of the CAR in vivo is dependent on the incorporation of a co-stimulatory domain into the fusion construct¹³⁵ (Fig. 1). Co-stimulatory cell surface receptors function with the TCR to generate a second signal for the activation of naive T cells¹³⁶. The first co-stimulatory domain to be incorporated into CARs for AIDs was from CD28, but further studies showed subtle benefits in using alternative co-stimulatory domains, such as one derived from the co-stimulatory molecule CD137 (4–1BB)^137,138. Efforts to enhance efficacy have evaluated innovative combinations of TCR-associated kinase domains¹³⁹ and screens of signalling domain libraries have pointed to potential advantages of heterologous domains for the function of CARs^140–142.

The CD28 and CD137 co-stimulatory components induce different levels of ‘tonic’ signals that drive rapid proliferation but can also cause T cell exhaustion^143,144. These co-stimulatory domains might also drive differential recruitment of Src-family kinases to these modules¹⁴⁵, and impact CRS¹⁴⁶ (Box 2). The prior observations in targeting malignant B cells might impact the use of CAR T cells in autoimmunity, although the criteria for the application in AIDs might require a revision of the paradigms established for oncology.

Notably, even in the absence of antigen, a low level of tonic signalling is mediated by the CAR CD3ζ subunit^147,148. The tonic signals are important for CAR T cell persistence, expansion and efficient synapse formation, as well as cytokine release, exhaustion and trafficking^149,150. Similarities and differences between TCR and CAR signalling have also motivated the development of synthetic HLA-independent TCR-like constructs that more closely approximate the binding sensitivity and regulatory mechanisms orchestrated by the TCR. Variously called HIT, STAR and TRuC receptors, these TCR-like constructs position the antibody VH and VL domains directly in frame with TCRα/β constant regions^151–153. This arrangement employs the principles of multimeric CD3 signalling, which more closely mirrors physiological TCR target engagement, enhances antigen-binding sensitivity and increases the capacity for serial target binding events. This innovative design has shown promise in early human trials for SLE¹⁵⁴.

Clearly, it is informative to examine the effectiveness and function of CARs (and related synthetic antigen receptors) in T cells based on principles of cell biology and antigen recognition by T lymphocytes. The persistence and in vivo efficacy of CAR T cells might be proportional to the efficacy and rate of antigen capture by T cells. T cells engage in trogocytosis, a process of acquiring membrane-bound subcellular particles from an APC or from cells expressing the cognate antigen¹⁵⁵. A higher rate of trogocytosis might be tied to T cell exhaustion and cooperative T cell killing¹⁵⁶. This was demonstrated for CD19, which is rapidly transferred by trogocytosis from B cells to CAR T cells, and the extent of antigen transfer directly impacts the efficacy and persistence of CAR T cells in vivo¹⁵⁶. Therefore, parameters of CAR binding to target antigen(s) will likely need to be tailored specifically to different autoimmune conditions in order to achieve deep, yet transient, B cell depletion and repertoire reset.

Hurdles to clinical implementation of CAR T cells

The implementation of CAR T cell therapies in AIDs introduces formidable clinical and organizational challenges. Chief among these are the requirements for specialized clinical expertise, advanced infrastructure and comprehensive safety oversight. Unlike haematologic malignancies, where haematologists have considerable familiarity with cell therapy protocols (for example, haematopoietic transplantation), in the autoimmune setting rheumatologists are likely to lack experience with CAR T cell therapies, including how to handle complications such as CRS and neurotoxicity¹⁵⁷ (Box 2). Therefore, the successful implementation of CAR T cell therapies in autoimmunity will depend on close collaboration and integration between oncologists and rheumatologists.

Even with the standardized protocols for autologous CAR T cell therapies used in multiple haematologic malignancies, the clinical deployment of these therapies remains largely restricted to specialized academic centres and is limited by complex logistics, safety considerations and cost. An expansion in manufacturing facilities has helped mitigate some of these bottlenecks, yet there remain significant barriers to the broader implementation of autologous CAR T cell therapies¹⁵⁸. Current manufacturing paradigms involve prolonged production timelines (typically 2–3 weeks), complex workflows (requiring cell purification, viral vector transduction and CAR T cell expansion), as well as the need for inpatient leukapheresis (the process used for white blood cell separation and collection), followed by reinfusion protocols. Autologous CAR T cell therapies thus remain burdened by high direct costs related to cell manufacture, patient hospitalization, lymphodepletion, bridge therapy between apheresis and infusion, and treatment of adverse events. These procedures also impose numerous expenses for patients due to the limited numbers of available centres and requirements of ongoing monitoring. A further major concern is the lack of scalability of autologous options for autoimmune indications.

An important hurdle for the scalability of CAR T cell therapies in autoimmune indications is the necessity for lymphodepletion of patients, which can pose treatment dilemmas. Lymphodepleting chemotherapy is an essential component of autologous CAR T cell therapies because it enhances CAR T cell engraftment and expansion. Aggressive lymphodepleting regimens (typically fludarabine and cyclophosphamide) induce transient but profound immunosuppression, which is justified in patients with life-threatening diseases. In most autoimmune indications, however, the risk–benefit assessment can shift significantly, especially for younger patients with non-life-threatening disease. Additional concerns include the time away from work or caregiving responsibilities, travel to specialized centres, psychological distress and fertility concerns (especially for younger women). Together, these issues can deter patient enrolment and reduce adherence to therapy. Furthermore, patients who are autoimmune often present with a history of immunosuppression and increased vulnerability to cytopenia, infections and adverse neurologic events. Such risks make lymphodepletion a greater concern for patients with AID than for those with cancer.

It is widely recognized that reducing or eliminating the need for lymphodepletion will broaden access to CAR T cell therapies in AIDs, as this will facilitate outpatient administration, increase patient safety and improve quality of life. The decision to offer CAR T cell therapy to patients who are autoimmune must carefully consider patient experiences, expectations and perspectives, as well as the burden of more traditional treatments against the opportunity to experience a durable reprieve from disease.

Future opportunities for CAR T cells

Driven by patient concerns, costs and a desire to expand patient access, various approaches are gaining attention and generating early positive results. In particular, the production of allogeneic approaches (Fig. 3), the engineering of alternative cell populations and the in vivo delivery of the CAR coding sequences (as DNA or RNA) to their intended target cells are attracting growing interest.

Allogeneic alternatives

There is growing interest in the development of ‘off-the-shelf’ allogeneic cellular therapies, where T cells are derived from healthy donor cells rather than from patients. Such CAR T cells must be genetically modified by gene editing to delete endogenous TCR and antigen-presenting function prior to their transfer into a patient. Allogeneic procedures aim to produce greater numbers of CAR T cells from a single or multiple donors. Following manufacture, allogeneic CAR T cells ought to minimize batch-to-batch variation, expand eligibility of CAR T treatment to include patients who cannot undergo apheresis and offer high-quality starting material. Allogeneic cryopreserved batches of T cells would be therefore immediately available for infusion, and for redosing if needed¹⁵⁹. Finally, the highly scalable manufacturing of allogeneic CAR T cells allows costs to be reduced and patient access expanded¹⁶⁰. However, allogeneic approaches present their own challenges: first, they expose patients to high risk of graft-versus-host disease and host-versus-graft reaction (immune rejection of allogeneic cells); second, they often require deeper lymphodepletion to allow engraftment¹⁶¹, which is associated with higher toxicities and a greater risk of infections; and third, similar to autologous therapies, they can induce CRS and immune-effector cell-associated neurotoxicity syndrome. The initial trials using allogeneic anti-CD19 CAR in SLE, myositis and SSc reported impressive progress in the clinical applications of this technically demanding approach^70,154.

Alternative T cell subsets

The task of depleting host B cells can also be implemented by using alternative immune cell sources instead of the traditional αβ T cells. Among different T cell subpopulations, natural killer (NK) cells, γδ T cells and regulatory T cells (T_reg cells) are some of the most promising options. CAR NK cells are more suitable for allogeneic applications, as they do not require HLA compatibility and exhibit favourable safety profiles¹⁶². Although they lack the highly diverse antigen recognition repertoire of T cells, NK cells have cytotoxic functions and can be activated through signalling that depends on CD3ζ (ref. 163). In the oncology setting, CAR NK cell therapies have shown favourable safety profiles but their efficacy and durability have been limited, with clinical benefits often lasting only a few months before relapse¹⁶⁴. CAR NK cells have difficulty reaching, infiltrating and persisting within tumours, yielding relatively short-term anticancer activity¹⁶². Fundamental biological features of NK cells, such as their limited lifespan and their target cell recognition, could pose challenges to achieving durable efficacy in the context of autoimmune disorders. Nonetheless, the field has rapidly advanced to overcome such challenges^165,166. At the beginning of 2025, 12 clinical trials using CAR NK cell therapies for AIDs were active. Of these, ten trials targeted CD19 and two used a bispecific CAR targeting CD19 and BCMA (Table 1 and Fig. 2). CAR NK cell therapy also holds promise in neuro-immunological disorders, such as MS, MG and neuromyelitis optica spectrum disorder (NMOSD), where NK cell dysfunction is linked to disease pathogenesis¹⁶⁷.

Similarly, the development of CAR γδ T cell therapies has experienced significant progress¹⁶⁸. Unlike conventional αβ T cells, γδ T cells have natural tissue-homing propensity, potentially achieving deeper cell depletion in tissues. Directed by antigen-specific CARs, the implementation of CAR γδ T cell therapies might benefit from the innate cytotoxicity of γδ T cells, their major histocompatibility complex (MHC)-independent antigen recognition and their reduced propensity to induce graft-versus-host disease. These attributes could enhance the safety and allogeneic applicability of CAR T cell therapies in the context of autoimmunity. Six clinical trials testing allogeneic CAR γδ T cells in AIDs are currently underway (Table 1).

An appealing alternative T cell subset is T_reg cells, which suppress excessive immune responses and maintain immune homeostasis. Engineering T_reg cells with CARs has emerged as a potentially promising therapeutic strategy for AIDs, where dysregulated immune activation can result in significant tissue damage^169,170. CAR T_reg cells can become activated by binding a single autoantigen but establish a broad suppressive environment. Two phase I clinical trials are testing CAR T_reg cells as therapeutics in hidradenitis suppurativa (HS) and RA (Table 1).

CAR-modified T_reg cells can effectively suppress autoimmune responses, but their stability under inflammatory conditions remains a critical concern¹⁷⁰. Another significant hurdle is that, unlike tumour-specific antigens, which are well defined in oncology, AIDs often involve multiple autoantigens and tissues, complicating the identification of suitable targets and increasing the risk of off-target effects^171,172. The manufacturing and scalability of CAR T_reg cells also present challenges. Autologous T_reg cells are often limited in number and might be dysfunctional in patients who are autoimmune, to which allogeneic T_reg cells offer a scalable solution. However, the suppressive function of T_reg cells can be inhibited in hostile microenvironments characterized by high inflammation or hypoxia, and CAR T_reg cells carry risks of reversion and polarization to helper T cells abetting graft-versus-host disease^173,174.Given that AIDs often affect diverse and hard to reach tissues, such as the central nervous system in MS or the joints in RA, engineering CAR T_reg cells to efficiently home to these sites is essential for therapeutic success¹⁷⁵. Conversely, over-suppression of the immune system by CAR T_reg cells could increase susceptibility to infection or malignancy, highlighting the need for fine-tuning local and systemic T_reg cell-mediated immunosuppression¹⁷⁵.

Despite these challenges, there are significant opportunities for innovation in CAR T_reg cell therapies. Leveraging synthetic biology approaches, such as designing CARs with dual signalling domains, could enhance the suppressive function of T_reg cells, even in inflammatory environments^176,177. Additionally, using epigenetic editing tools to stabilize T_reg cells with a regulatory phenotype represents a promising direction for future research¹⁶⁹. The development of allogeneic CAR T_reg platforms using gene-edited T_reg cells could help overcome the limitations associated with autologous cells¹⁷³. Temporal engineering of CAR T_reg cells, wherein cells are activated only during autoimmune flares, could mitigate the risks associated with long-term immune suppression. Lastly, designing synthetic circuits that respond dynamically to inflammatory signals could make CAR T_reg cells more adaptable and safer¹⁷⁷.

In vivo CAR T cells

Most CAR T cell products are generated using ex vivo manufacturing, which is efficient and well validated but has multiple drawbacks, as discussed above¹⁷⁸. To circumvent the complexity, costs, inefficiencies and safety risks of ex vivo generation of CAR T cell therapies, the field is advancing innovative in vivo strategies that have the potential to avoid lymphodepletion and allow better control over CAR kinetics. In these approaches, gene therapy vectors encoding a CAR are administered to a patient and become transduced into T cells in vivo. For example, implanted biomaterials have shown promise for CAR T cell engineering in vivo¹⁷⁹. In vivo alternatives include both viral and non-viral approaches for CAR T cell transduction¹⁸⁰. Preclinical studies have demonstrated the feasibility of in vivo CAR delivery by viral vectors such as lentiviral and adenovirus-associated vectors¹⁸¹. In addition, there are ongoing clinical trials (Table 1) using lentiviral vectors to engineer CAR T cells in vivo. However, viral vectors face multiple limitations, including restrictions on insert size dictated by viral capsids, mutagenic and oncogenic risk, the presence or induction of immunogenicity, extended manufacturing procedures and elevated costs¹⁷⁸.

Among the most promising non-viral in vivo CAR T cell technologies are lipid nanoparticles (LNPs) that target the modified mRNA encoding the CARs to T cells^182,183 (Fig. 3). This approach has been successful in vaccine production¹⁸⁴ and shows potential for industrial scalability¹⁸⁵. LNPs not only circumvent the size constraints of viral vectors but also overcome the risk of insertional mutagenesis, while reducing manufacturing costs¹⁷⁸. Although generally considered less immunogenic than viral vectors, LNPs still might elicit immune reactions. However, the optimization of their lipid composition and fine-tuning of nucleotide modifications in the mRNA cargo can reduce their immunogenicity¹⁸⁶.

Although mRNA delivery allows only transient CAR expression, this might be sufficient and even desirable in several AIDs, allowing reconstitution of a reset immune system and retreatment if disease recurs. Despite the potential of in vivo CAR T cell engineering, several drawbacks need to be addressed, such as off-target CAR delivery and infusion reactions upon systemic administration of nanoparticles¹⁸⁷. Improved formulation parameters of the LNP can enhance mRNA integrity and increase transfection efficiency¹⁸⁰. Theoretically, other RNA species such as self-amplifying or circular RNAs could prolong the expression of genes of interest¹⁸⁰ and enhance the therapeutic benefit of a single dose of RNA.

An elegant workaround to the in vivo CAR is available through innovative mRNA autologous CAR T cell platforms, in which mRNA rather than DNA is used to encode the CAR in T cells that then can be administered to the patient without lymphodepletion. In a phase II randomized, double-blind, placebo-controlled trial for patients with MG, one infusion of an investigational mRNA anti-BCMA autologous CAR T cell therapy (Descartes-08), administered in an outpatient setting, led to deep and durable responses through 12 months of follow up⁷⁶. This programme was recently advanced to phase III.

Concluding remarks

It is increasingly accepted that CAR T cell therapy holds advantages over the current standard of care in the treatment of AIDs, which has changed little in decades and remains mostly based on long-term immunosuppressive therapies. As described here, depletion of B cells with CARs to CD19 is only one of several available approaches to disrupt the progression of autoimmunity. Experiences with CAR T cell therapy indicate that if the disruption is sufficiently profound, the immune system can gain the capacity to regulate itself and reestablish self-tolerance. The added prospect that CAR T cell therapy might require only periodic medications or even lead to drug-free remission has therefore generated tangible enthusiasm. However, it should be acknowledged that the underlying predisposition towards immune self-reactivity is not eliminated by the treatment. Thus, it is possible that autoimmunity could recur later due to the precarious balance between tolerance and autoimmunity that is inherent in the immune system.