Remission of severe myasthenia gravis after autologous stem cell transplantation

Abstract

Objective

Myasthenia gravis (MG) is an autoantibody‐mediated neuromuscular junction disorder involving the acetylcholine receptors on the motor endplate. The safety and response to high‐dose chemotherapy (HDIT) and autologous hematopoietic cell transplantation (HCT) were assessed in a patient with severe refractory MG.

Methods

As part of a pilot study of HDIT/HCT for patients with treatment‐resistant autoimmune neurological disorders, a patient with severe refractory MG underwent treatment. After mobilization of hematopoietic stem cells with rituximab, prednisone, and G‐CSF, the patient had HDIT consisting of carmustine, etoposide, cytarabine, melphalan, and rabbit antithymocyte globulin, followed by autologous HCT. The effect of treatment on the autoantibody to the acetylcholine receptor (AChR) was assessed.

Results

The patient had been diagnosed with AChR antibody‐positive MG 14 years before HDIT/HCT and had failed thymectomy, therapeutic plasma exchange, and multiple immunomodulatory agents. The Myasthenia Gravis Foundation of America (MGFA) clinical classification was IVb before HDIT/HCT. She tolerated HDIT/HCT well and started to improve clinically within days of treatment. At both 1 and 2 years after HDIT/HCT, patients remained symptom‐free. After HDIT/HCT, AChR‐binding autoantibodies persisted, and the relative frequency of immune cell subtypes shifted.

Interpretation

HDIT/HCT induced a complete response of disease activity in a patient with severe refractory MG. This response may suggest that a cell‐mediated etiology may be a significant contributing factor in refractory MG cases. A phase 2 clinical trial is warranted to establish if HDIT/HCT can be an effective therapy for severe refractory MG and to gain a further understanding of disease pathogenesis.

Introduction

Myasthenia gravis (MG) is considered a prototypical antibody‐mediated neurological autoimmune disorder. ¹ It affects the neuromuscular junction causing, weakness of the skeletal muscles and motor fatigue. Approximately 80% of generalized MG patients have autoantibodies to the acetylcholine receptor (AChR). ² , ³ , ⁴ Other autoantibodies associated with the disease include muscle‐specific kinase (MuSK) and lipoprotein receptor‐related protein 4 (LRP4), both associated with the neuromuscular junction postsynaptic membrane. Additionally, 10–20% of patients may be seronegative. ² , ³ , ⁴ , ⁵ , ⁶ Although most patients will respond to currently available therapies, 10–20% of patients may have refractory MG, and an even smaller percentage of cases will be severely debilitated by the disease. ⁷ , ⁸ , ⁹

High‐dose immunosuppressive therapy (HDIT) and autologous hematopoietic cell transplantation (HCT) have proven to be viable treatment for severe autoimmune diseases (AID) over the last two decades. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ After HCT, dominant TCR clones in CD4⁺ T cells present before treatment were significantly depleted, and following reconstitution, patients had largely developed a new T‐cell repertoire. ¹⁵ , ¹⁶ This may explain the long‐term remissions that have been observed after HCT. The experience is still very limited for patients with MG undergoing HDIT and autologous HCT, especially for patients with very severe disease. ¹⁷ , ¹⁸ , ¹⁹

We hypothesized that patients with severe, refractory MG to conventional therapy would respond to HDIT and autologous HCT. We describe here our experience treating a patient with severe refractory MG on a clinical trial with 2 years of follow‐up after autologous HCT.

Subjects/Materials and Methods

Subject

The patient was enrolled in the study in January 2020. The study was approved by the WCG Institutional Review Board (IRB) for the Colorado Blood Cancer Institute (CBCI) and Presbyterian/St Luke's Medical Center (PSL), as well as by the central IRB at the Fred Hutchinson Cancer Research Center (FHCRC). The patient was registered for the study after providing written informed consent to confirm that they were fully aware of the investigational purposes of the clinical trial in accordance with the Declaration of Helsinki. The patient was evaluated and transplanted at CBCI and PSL. Research samples were cryopreserved at FHCRC, and the laboratory evaluation was performed at Yale University. The clinical trial was registered at clinicaltrials.gov (NCT00716066) and includes patients with a variety of rare and severe autoimmune neurological diseases.

Study design

This pilot study was designed to assess the safety and potential efficacy of HDIT and autologous HCT for patients with severe autoimmune neurological diseases. To be eligible for the study, the patient was required to be less than 71 years of age. To satisfy the diagnostic criteria for MG, the patient was required to have a positive serological test for AChR antibody and an abnormal electrodiagnostic study consistent with a diagnosis of MG. Patients were required to have active disease despite treatment or a history of recurrent myasthenic crises. Disease severity before and after autologous HCT was assessed using the Myasthenia Gravis Foundation of America (MGFA) clinical classification, the Myasthenia Gravis Composite (MGC) score, and the Myasthenia Gravis Activities of Daily Living (MG‐ADL) score.

Treatment and supportive care

Mobilization and collection of stem cells

On Day 1 of mobilization, the patient received rituximab (375 mg/m²) and started prednisone (1 mg/kg/day) for 10 days. G‐CSF (16 mcg/kg/day) was administered subcutaneously on day 2 and finished on the last day of collection of stem cells. The first day of stem cell collection was Day 5. The targeted CD34⁺ cell count was ≥4.0 × 10⁶ cells/kg. The unmanipulated G‐CSF–mobilized peripheral blood stem cell (PBSC) graft was cryopreserved.

Transplantation

The HDIT regimen was carmustine (BCNU) 300 mg/m² on Day −6 of autologous HCT, etoposide 100 mg/m² twice daily, cytarabine 100 mg/m² twice daily from day −5 to −2, and melphalan 140 mg/m² on Day −1 (BEAM). ¹¹ Rabbit antithymocyte globulin (2.5 mg/kg/d) was administered intravenously on Days −2 and − 1. The patient was given methylprednisolone (1 mg/kg IV) with each rATG dose, with a repeat dose of 1 mg/kg 12 h later. On Day 0, the autologous PBSC graft was thawed and infused. G‐CSF (5 mcg/kg/day) was administered from day +7 until the recovery of blood counts. Prednisone was administered (0.5 mg/kg/d) from days +7 to +21 and then tapered over 2 weeks to reduce the risk of engraftment syndrome. ¹¹ The patient was hospitalized from the start of HDIT until there was neutrophil engraftment and resolution of major toxicities. Infection prophylaxis included cefdinir for severe neutropenia, trimethoprim‐sulfamethoxazole for Pneumocystis jiroveci (given for 1 year), fluconazole for Candida (75 days), and acyclovir for herpes simplex virus (HSV) and varicella‐zoster virus (1 year). Monitoring for cytomegalovirus (CMV) reactivation continued until 1 year after HDIT. Surveillance monitoring for Epstein–Barr virus reactivation began on Day +14 post‐transplant.

Evaluation of outcomes

The primary objective of the study was to evaluate the safety of high‐dose carmustine, etoposide, cytarabine, melphalan and rabbit antithymocyte globulin as a HDIT regimen in patients with severe, refractory neurological autoimmune disease. The secondary objective was to evaluate disease responses and the duration of response to HDIT and autologous HCT.

The day of neutrophil engraftment was defined as the first of three consecutive days after HCT when an absolute count of greater than 0.5 × 10⁹/L (500/mcL) had been achieved. The day of platelet engraftment was defined as the first of three consecutive days when a count of 20 × 10⁹/L (20,000/mcL) had been sustained or the first of 2 days in which the platelet count increased on the second day without transfusions. Regimen‐related toxicities were reported utilizing the NCI Common Terminology Criteria for Adverse Events (CTCAE) modified for blood and marrow transplantation.

Scheduled evaluations of the patient were performed before mobilization, then at 3 months and annually after HCT. These evaluations included a medical history, physical examination, complete blood count, serum chemistries, autoantibodies (anti‐AChR), thyroid function tests, electrodiagnostic studies, and standard measurements used to evaluate patients for HCT.

Serological autoantibody assays

The clustered AChR autoantibody‐binding live cell‐based assay (CBA) was performed as previously described ⁵ , ²⁰ using serial dilutions of serum, plasma (1:20, 1:50, 1:150, 1:450, 1:1350, 1:4050) or isolated Ig (200, 100, 50, 25, 12.5, 6.25 μg/ml), which was isolated using a Melon Gel IgG Spin Purification Kit (45206; ThermoFisher), according to the manufacturer's instructions. AChR autoantibody‐mediated complement fixation was evaluated with a live CBA, as we have previously described. ²¹ , ²²

Immunophenotyping

Four flow cytometry panels (Table S1) were used to examine the frequencies of B, T (two panels), and NK immune cell subsets in unstimulated cryopreserved peripheral blood mononuclear cells (PBMC). PBMC samples consisted of two pre‐ (1136 and 53 days pre‐HCT) and two post‐ (368 and 642 days post‐HCT) HCT samples, as well as four age‐ and sex‐matched HD samples from the Yale Neuromuscular Biorepository (female, mean age 28.3). All samples analyzed were verified to have PBMC viability greater than 80%, as assessed by a LIVE/DEAD™ Fixable Dead Cell Stain Kit (Thermo Fisher). Briefly, samples were thawed and incubated in the dark at room temperature with viability dye for 20 min. Following washing, PBMCs were split into approximately 5 × 10⁵ cells for each panel, then incubated with panel‐specific antibody cocktails at 4°C for 45 min, and then analyzed with a BD LSRFortessa flow cytometer and FACSDiva software. UltraComp eBeads (Invitrogen) were utilized to correct fluorescent spillover. Analyses were performed using FlowJo software and GraphPad Prism.

Results

Case report

The patient was a 33‐year‐old female with a past medical history of a splenectomy for a spleen‐associated hemangioma who presented for autologous HCT in 2020. She was diagnosed with MG 14 years before autologous HCT based on abnormal repetitive nerve stimulation (RNS) testing at 3 Hz showing a significant decremental response and positive AChR autoantibodies. She was initially treated with prednisone and pyridostigmine, followed by an elective thymectomy 1 year later. She was then started on mycophenolate mofetil (MMF), which improved her symptoms but continued to cause occasional flares for several years.

MMF was later tapered off in the setting of a planned pregnancy, delivering 5 years before autologous HCT. Postpartum, her MG worsened, and despite re‐initiation of MMF, prednisone, and IVIg, her disease was not well controlled. Further optimization of her treatment regimen included the addition of therapeutic plasma exchange (TPE) and rituximab, with only minimal benefit. Given persistent symptoms, she started eculizumab 3 years before autologous HCT, to which she only had a brief and modest response. She was then treated with cyclophosphamide, which was reported as helping modestly but was not continued due to side‐effect concerns. She remained on pyridostigmine, low‐dose prednisone, TPE, and IVIG every 2 weeks. During the 3 years before autologous HCT, she had approximately 10 hospitalizations requiring ICU admission, intubation three times, and non‐invasive positive pressure ventilation (NIPPV) the other times. Given the refractoriness of her disease, she was referred to the transplant program at CBCI and PSL for HDIT and autologous HCT. Pre‐transplant, she had a MGFA Classification of IVb. The MGC score was 38, and the MG‐ADL score was 17 (Table 1).

Table 1.

Clinical outcomes.

	Pre‐transplant	+4 months	+10 months	+21 months
MGFA Clinical Classification a	IVb	0	0	0
MGC Score b	38	0	0	0
MG‐ADL Score c	17	1	0	0

^a The MGFA Clinical Classification is used to identify subgroups of MG patients that share clinical features, including severity of disease and distribution of weakness. It ranges from 0 indicating remission (symptom‐free state) and 1 indicating ocular disease to 5, which is an MG crisis. MGFA severity classes II–IV can be designated as a or b, depending on whether the weakness is predominantly limb or axial weakness (a) or oropharyngeal or respiratory weakness (b).

^b The MGC score is a validated, patient‐ and physician‐reported 10‐item assessment tool for evaluating the symptoms and signs of MG. The total score ranges from 0 to 50, with higher scores indicating a greater impact of MG on functional activities.

^c The MG‐ADL is a patient‐reported, validated, disease‐specific instrument that assesses MG symptoms and daily activities commonly affected by the disease. The MG‐ADL has been used as the primary outcome measure in multiple MG clinical trials. It contains eight items and is scored 0–24, with higher scores representing a greater burden of disease.

When she presented for autologous HCT, she had fatigue, double vision, ptosis, weakness in her arms and legs, and was wheelchair‐bound. She spoke in a whisper and had difficulty breathing and swallowing. She required a feeding tube. Laboratory values at the time were normal except for mild anemia. The MG‐specific clinical reference laboratory tests prior to HDIT/HCT revealed an elevated AChR‐binding antibody at 1.92 nmol/L (AChR‐blocking antibody <15% and AChR‐modulating antibody 48%). RNS at 3 Hz (bilateral spinal accessory and the right radial nerves) showed significant decremental responses consistent with a diagnosis of MG. Pulmonary function tests revealed an FEV1 of 55–61% predicted and a DLCO adjusted for hemoglobin of 77%. The FVC was 82% predicted and increased to 90% predicted with bronchodilators. An echocardiogram was unremarkable, with a left ventricular ejection fraction of 75%.

A total of 4.66 × 10 (6) CD34⁺ cells/kg was collected after mobilization of stem cells. Given her worsening clinical symptoms, she again received TPE prior to starting HDIT.

The patient proceeded with BEAM, requiring emergent TPE during the treatment course due to worsening respiratory symptoms that required NIPPV for support. BEAM was held for 1 day during TPE. The stem cell rescue occurred on Day 0. By Day +3, she was no longer whispering but had a stronger, clearer voice. She achieved neutrophil engraftment on Day +10 after HCT. Recovery of platelet counts occurred by Day +12.

She had progressive weakness on Day +14, including worsening respiratory status as evidenced by a small drop in her negative inspiratory force (NIF). She required brief support with NIPPV and received another TPE. She remained on pyridostigmine during her treatment. She received prednisone 10 mg daily prior to the HDIT/HCT and continued this during the treatment, increasing the dose on Day +7 to 30 mg daily (0.5 mg/kg/day). She remained on this dose through Day +21 of the HCT and then tapered per protocol.

Her venous blood gas and spirometry remained stable throughout her treatment. She had intermittent nocturnal hypoxemia but, by discharge, was on room air. She was discharged from the hospital on Day +16.

Before discharge from the hospital, she experienced a neutropenic fever (grade 3), a rash (grade 2), and EBV reactivation, which did not require treatment (grade 1). These were resolved by discharge or soon after discharge. No adverse events were observed after the immediate post‐transplant period.

On her first outpatient clinic visit on Day +19, she reported improved speech and control of facial muscles. Her legs were stronger, allowing her to use her walker more frequently rather than a wheelchair. Ptosis was completely resolved by Day +21. She noted an improved upward gaze and an improved gag reflex. She had occasional nausea, vomiting, and dry mouth, which were attributed to HDIT.

She tapered off prednisone completely by Day +36. She remained on pyridostigmine. At Day +35, post‐transplant RNS testing was repeated, which showed resolution of the previously seen decrement of the bilateral spinal accessory and the right radial nerves. TPE and IVIg treatments were discontinued.

At 4 months post‐transplant, the patient had a MGFA clinical classification score of 0. The MGC score was 0, and the MG‐ADL score was 1 (Table 1).

At 12 months post‐transplant, RNS testing was normal. Clinical reference laboratory tests showed that the AChR‐binding antibody was 0.93 nmol/L (AChR‐blocking antibody 31% and AChR‐modulating antibody <12%). T cell testing revealed a low CD4⁺ T cell fraction and CD4/CD8 ratio and elevated absolute CD8⁺ T cell and NK cell counts.

The neurological examination was normal at 1 year. The MGFA clinical classification, the MGC score, and the MG‐ADL score, which were obtained at 10 months after HDIT/HCT, were all 0 (Table 1). Post‐transplant, her AChR‐binding antibody level was not significantly changed with complete resolution of symptoms.

At 24 months post‐transplant, the patient was neurologically normal, and the RNS testing remained normal. The MGFA clinical classification, the MGC score, and the MG‐ADL score obtained at 21 months after HDIT/HCT were all 0 (Table 1). She discontinued pyridostigmine 26 months after HDIT/HCT.

AChR‐binding autoantibodies persist post‐transplant

To further evaluate autoantibody binding, we performed research laboratory‐based serological screening using a live clustered AChR CBA. AChR autoantibodies were present in the serum before (mean ΔMFI_1:20 634) and after HDIT/HCT (mean ΔMFI_1:20 1080) when compared to HD (mean ΔMFI_1:20 −11) and MG controls (mean ΔMFI_1:20 1556). Serum‐isolated Ig also exhibited AChR binding, pre‐ (mean ΔMFI_200ug/mL 496) and post‐ (mean ΔMFI_200ug/mL 437)HDIT/HCT compared to HD (mean ΔMFI_200ug/mL −18) and MG (mean ΔMFI_200ug/mL 2220) controls (Fig. 1A,B).

Figure 1.

Acetylcholine receptor (AChR)‐binding autoantibodies persist post‐hematopoietic cell transplantation (HCT), though autoantibody‐mediated complement activity diminishes. Live CBA using transiently AChR‐transfected HEK293T of (A) serum and (B) isolated Ig to detect the presence of AChR autoantibodies in three pre‐HCT and two post‐HCT serum samples. (C) Representative flow cytometry plots depict MAC formation, indicative of AChR autoantibody‐mediated complement activity, upon incubation of pre‐HCT serum with AChR‐expressing cells in a modified CBA. (D) AChR autoantibody‐mediated complement activity measured in pre‐HCT and post‐HCT samples, HD, and an AChR‐positive MG patient serum known to mediate autoantibody‐mediated complement activity. The dotted line represents background (NHS only) complement activity. Each point represents the mean of triplicate experimental conditions. Extended results for the assay data are shown in Table S2.

We then examined the potential pathogenicity of these autoantibodies by testing their ability to activate the complement pathway using a modified AChR CBA (Figure 1C,D). ²¹ , ²² Low AChR autoantibody‐mediated complement activity was detected in the pre‐transplant samples (median MFI 144) compared to the MG controls (median MFI 405). The post‐transplant samples showed levels of AChR autoantibody‐mediated complement that were similar to those of HD controls (median MFI_post‐AHCT 97 vs. MFI_HD 93).

Immunophenotyping

Analyses of B and T lymphocyte phenotypes were performed with peripheral blood mononuclear cell (PBMC) samples obtained immediately pre‐transplant and at two post‐transplant time points (Table 2). B‐cell subset measurements revealed a sustained naïve population (CD27⁻IgD⁻) to the exclusion of the experienced (CD27⁻) compartment. In comparison to the pre‐transplant sample, we observed a decrease in T (total), CD4⁺ T, Treg, and memory CD8⁺ T cells. We found an elevation in the total CD8⁺ T cells at both post‐transplant time points. NK cell and subset frequencies were analyzed in the pre‐ and post‐transplant samples and four HDs based on CD16 and CD56 expression (Fig. 2). Total and immunomodulatory (CD56^hiCD16⁻) NK cell frequencies were not different between the HD and patient‐derived samples. However, mean pre‐transplant frequencies of cytotoxic NK cells (CD56^loCD16^hi) were 30.8% lower (3.3xSD_HD) than those found in HD; these subset frequencies returned to levels similar to those in HD in the post‐transplant period and were maintained at that level for over 2 years.

Table 2.

B and T lymphocyte frequencies.

Sample	−53 days	+368 days	+642 days	Marker definition
T cells	49.4%	43.2%	33.7%	CD3+
CD4 T cells*	49.40%	38.80%	30.4%	CD3+ CD4+ CD8−
CD8 T cells*	41.70%	47.90%	58.3%	CD3+ CD8+ CD4−
Memory CD4 T cells^	65.23%	68.07%	38.85%	CD3+ CD4+ CD8− CD45RA−
Activated CD4 T Cells^	1%	0.63%	0.78%	CD3+ CD4+ CD8− HLA‐DR+ CD38+
Memory CD8 T Cells~	21.1%	18.4%	6.47%	CD3+ CD4− CD8+ CD45RA−
Activated CD8 T Cells~	0.32%	0.57%	0.15%	CD3+ CD4− CD8+ HLA‐DR+ CD38+
Regulatory T cells^	6.38%	5.97%	3.53%	CD3+ CD4+ CD25+ CD127−
B cells	3.08%	8.66%	1.44%	CD3− CD14− CD19+
Naïve B cells+	89.50%	89.40%	85.70%	CD3− CD14− CD19+ IgD+ CD27−
Memory B cells+	3.43%	1.96%	4.32%	CD3− CD14− CD19+ IgD− CD27+
Double negative B cells+	4.01%	6.64%	9.01%	CD3− CD14− CD19+ IgD− CD27−
Activated B cells+	1.33%	1.86%	3.27%	CD3− CD14− CD19+ IgD− CD71+
Antibody secreting B cells+	0.22%	0.15%	0.46%	CD3− CD14− CD19+ CD27++ CD38++

Gated out of live lymphocyte singlets (d = days, *frequency out of total T cells, ⁺frequency out of total B cells, ^frequency out of CD4⁺ T Cells, ~frequency out of CD8⁺ T Cells).

Figure 2.

Skewed natural killer (NK) cell subset frequency is normalized following hematopoietic cell transplantation (HCT). Reduced cytotoxic NK cell frequency was observed prior to HCT and returned to HD levels post‐HCT. (A) Representative flow cytometry plots depict NK cell subsets based on CD16 and CD56 expression. (B) Frequencies of NK cell subsets as a percentage of the parent population are shown. NK cells are gated as CD56⁺CD3⁻CD14⁻CD19⁻ from the singlet lymphocyte population; the remaining NK cell subsets are displayed as the frequency of the NK cell gate. Each dot for pre‐ and post‐HCT represents a different time point (n = 2), and each dot for HD represents an individual donor (n = 4); bars show the mean and SEM.

Discussion

In this case report, a patient with refractory severe MG underwent HDIT and autologous HCT with complete neurological recovery. Improvement was first noted within days after HDIT/HCT, and she progressively improved in subsequent months. Abnormal decremental response on RNS testing resolved 30 days after transplant. Because of the patient's severe neurological disability, she was at increased risk of morbidity and transplant‐related mortality, but she recovered without major complications.

HDIT and autologous HCT induce remissions and prolong overall survival and event‐free survival in patients with systemic sclerosis. ¹³ , ¹⁴ , ²³ For multiple sclerosis (MS), event‐free survival was 70% at >5 years after transplant. ¹¹ A randomized clinical trial in MS patients demonstrated improved time to disease progression after HDIT and autologous HCT compared to conventional therapy. ¹² Guidelines have been developed for the use of HDIT and autologous HCT for relapsing/remitting MS based on the results of the clinical trials reported to date. ²⁴ , ²⁵ , ²⁶ There is still limited experience with HDIT and autologous HCT for MG. In 2016, a case series from a single institution reported seven patients with severe MG at a median follow‐up of 40 months after transplant, and all were reported to have completely stable remissions with no recurrence of disease. ¹⁷ One of these patients continued low‐dose pyridostigmine for 5 years after HDIT/HCT. In 2017, a case was reported of a patient with severe MG who underwent HDIT and autologous HCT 38 years after diagnosis. ¹⁸ She was off all treatment by 10 months after HDIT/HCT and had mild residual symptoms related to persistent disease. The anti‐AChR autoantibodies significantly decreased after transplant but persisted at low levels. There was a report of a pediatric case of severe MG that showed marked clinical improvement but also had persistence of the anti‐AChR autoantibodies after allogeneic HCT. ²⁷ The persistence of the autoantibodies after allogeneic HCT may have occurred because the patient had mixed chimerism post‐transplant. In 2019, another case of a patient undergoing HDIT/HCT for severe MG remained in remission 65 months after transplant. ¹⁹ In 2023, a small case series reported on three patients with anti‐MuSK MG treated with HDIT/HCT, all of whom became symptom‐free from MG with a tolerable side effect profile. ²⁸ Here we report a more detailed assessment of AChR autoantibodies and other immunological parameters than previously reported after HDIT/HCT for MG.

Although MG is considered an antibody‐mediated autoimmune disorder, this patient did not exhibit a significant change in AChR autoantibody level after HDIT/HCT. Only two other transplant studies included autoantibody quantification and reported a persistent AChR autoantibody level post‐transplant. ¹⁸ , ²⁷ Additionally, AChR autoantibody‐mediated complement activity was very low prior to HDIT/HCT and not measurable after. This finding suggests that autoantibody‐mediated complement activation may not be a major contributor to pathology in this patient, a possibility that is in alignment with the patient's poor response to eculizumab. The proportions of blocking and modulating autoantibodies shifted after the transplant. However, given that autoantibody titers and their associated pathogenic mechanisms did not show major changes after the transplant, it is challenging to associate autoantibody‐mediated mechanisms with the marked clinical improvement. It is also possible that the serum autoantibodies we investigated do not reflect pathogenic activity at the neuromuscular junction or that the autoantibody repertoire was altered in a manner that was not measured in the assays used in this study. Thus, to explore other possibilities, we turned to investigating changes in immune cell frequencies.

Through immunophenotyping, we observed a maintenance of the naïve B‐cell population, which may be a consequence of prior rituximab treatment, and no other remarkable changes in B‐cell subset frequencies. However, our data recapitulate prior findings of elevated CD8⁺ T and reduced CD4⁺ T cell frequencies following HDIT and autologous HCT for systemic sclerosis and multiple sclerosis. ²⁹ These changes may have included increased proportions of regulatory phenotypes associated with the clinical improvement. However, specimen availability limited the possibility of investigating these subsets further. A reduction in cytotoxic NK cell frequency was observed before HDIT/HCT, which then increased to levels comparable to the HD cohort after HDIT/HCT. Reductions in NK cell frequencies in other autoantibody‐mediated diseases, particularly neuromyelitis optica spectrum disorder and systemic lupus erythematosus, have previously been reported; however, further study is required to delineate whether this phenomenon is recapitulated in MG. ³⁰ , ³¹ , ³² These findings suggest that NK cells may contribute to MG pathology, as the reduction in peripheral populations may signify egress to areas of pathology. Investigating this possibility would ideally include the evaluation of immune cells that have localized at the MG neuromuscular junction, which are thought to be infrequent but have not been fully explored. It is also possible that these findings reflect antibody‐dependent or antibody‐independent cellular cytotoxicity that occurred before HDIT/HCT. ³³ , ³⁴ Collectively, these findings provide support for further investigation into the role of both T and NK cells in severe MG resistant to conventional treatments and the therapeutic mechanisms of HDIT/HCT.

Other studies of immune reconstitution after transplantation demonstrated that HDIT and autologous HCT may “reset” the immune system. ¹⁵ , ¹⁶ HDIT and autologous HCT have distinct effects on CD4⁺ and CD8⁺ T‐cell repertoires. In CD4⁺ T cells, dominant TCR clones present before treatment were depleted, and following reconstitution, patients largely developed a new CD4⁺ T‐cell repertoire. In contrast, dominant CD8⁺ clones were not effectively removed, and the reconstituted CD8⁺ T‐cell repertoire was largely created by clonal expansion of cells present before treatment. ¹⁵ In a follow‐up study, it was observed that HDIT and autologous HCT removed the majority of T cells existing in the intrathecal compartment of patients with active RRMS, likely including pathogenic T cells, and induced the generation of a newly differentiated immune repertoire that persisted in both intrathecal and peripheral blood compartments. ¹⁶ This effect of transplantation on the immune repertoire may be the most relevant for understanding the prolonged remissions observed in patients with MG, given the well‐established role that CD4⁺ T cells play in autoantibody production. ³⁵ , ³⁶

In summary, this report describes a patient with severe refractory MG with a sustained complete response to high‐dose therapy and autologous HCT. The persistence of the AChR autoantibody levels along with early clinical improvement after transplantation suggests a cell‐based disease mechanism that responded to high‐dose therapy. Based on these encouraging early findings for transplantation of refractory MG and other autoimmune diseases, a phase 2 clinical trial, coupled with mechanistic studies, is warranted to assess safety and the frequency and durability of responses.

Author Contributions

MIS: acquisition of data and drafting the manuscript. SSY: design of the mechanistic studies, data acquisition, analysis, drafting the manuscript, and preparation of the figures. KCO: design of the mechanistic studies, data acquisition, analysis, and drafting the manuscript. RJN: data acquisition, analysis, and drafting the manuscript. MCP: data acquisition, analysis, and preparation of the figures. AHO: data acquisition, analysis, and preparation of the figures. CR: data acquisition. MP: data acquisition. PAM: drafting the manuscript. MLP: data acquisition (study neurologist). MTT: drafting the manuscript. JDB: study concept and design and drafting the manuscript. RAN: study concept and design, data acquisition, analysis, drafting the manuscript. GEG: study concept and design and drafting the manuscript.

Funding Information

Dr. Kevin C. O'Connor is supported by the National Institute of Allergy and Infectious Diseases of the NIH under award numbers R01‐AI114780 and R21‐AI164590; through a pilot award (MGNet Pilot (21‐M99) provided through the Rare Diseases Clinical Research Consortia of the NIH and MGNet (award number U54‐NS115054), and by a High Impact Clinical Research and Scientific Pilot Project award from the Myasthenia Gravis Foundation of American (MGFA). Dr. Richard J. Nowak is supported by the NIH (award number U54‐NS115054).

Conflict of Interest

Dr. Kevin C. O'Connor has received research support from, Alexion, (now AstraZeneca) and argenx, and is an equity shareholder of Cabaletta Bio.

Dr. Nowak reports research support from the National Institutes of Health, Genentech, Inc., Alexion Pharmaceuticals, Inc., argenx, Annexon Biosciences, Inc., Ra Pharmaceuticals, Inc. (now UCB S.A.), the Myasthenia Gravis Foundation of America, Inc., Momenta Pharmaceuticals, Inc., Immunovant, Inc., Grifols, S.A., and Viela Bio, Inc. (Horizon Therapeutics plc). Dr. Nowak has also served as a consultant and advisor for Alexion Pharmaceuticals, Inc., argenx, Cabaletta Bio, Inc., COUR, CSL Behring, Grifols, S.A., Ra Pharmaceuticals, Inc. (now UCB S.A.), Immunovant, Inc., Momenta Pharmaceuticals, Inc., and Viela Bio, Inc. (Horizon Therapeutics plc).

Supporting information

Table S1

Funding Statement

This work was funded by Myasthenia Gravis Foundation of America ; National Institute of Allergy and Infectious Diseases of the NIH grants R01‐AI114780 and R21‐AI164590; Rare Diseases Clinical Research Network grant U54‐NS115054.