Abatacept is useful in autoimmune cytopenia with immunopathologic manifestations caused by CTLA-4 defects

TO THE EDITOR:

Autoimmune diseases can reveal underlying primary immune deficiencies (PIDs), among which are cytotoxic T-lymphocyte–associated antigen 4 (CTLA-4) pathway variants. From 2015, this pathway has become a classical clinical entity combining autoimmune cytopenia (AIC) and various immunopathologic manifestations (IMs).^1,2 The role of CTLA-4 in controlling lymphocyte activation can be disrupted in case of dominant heterozygous CTLA-4 variants,^3,4 recessive homozygous mutations, or deletions of lipopolysaccharide-responsive beige-like anchor (LRBA)^5,6 or DEF6 variants.⁷ Abatacept, a CTLA-4 fusion protein labeled in rheumatoid arthritis,⁸ is a logical targeted option for those patients.

In 2019, a preliminary study of the French OBS’CEREVANCE cohort of children with AIC identified CTLA-4 pathway pathogenic variants in 10 of 80 cases of pediatric-onset Evans syndrome (pES).⁹ pES is now recognized as a severe long-lasting disease: numerous IMs appearing with increasing age in 78% of cases; high burden of second-line treatments to control severe bleeding or hemolytic flares or organ-specific IMs; and mortality rate reaching 10%.^10,11 Targeted therapies are urgently warranted to improve the outcome in those patients.

In May 2020, 19 of 1613 children from this national cohort were identified with CTLA-4 defects, and 9 of them were treated with abatacept. We aimed to describe the impact of this treatment on the course of their disease.

This was a national observational study. Of note, patients younger than 18 years of age with AIC are routinely included in real time in the OBS’CEREVANCE cohort, prospectively followed, and screened for an underlying diagnosis of IM and PID.^9,11,12 Written informed consent was obtained from parents and patients. Patients with underlying CTLA-4 pathway pathogenic variants who were treated with abatacept were analyzed in May 2021 with a more than 1-year follow-up for all. End points were the effect on AIC and IM according to standard definitions and the burden of further second-line treatments. Adverse events and their grades were recorded (CTCAE v5.0). Continuous variables were described as the median (range), and categorical variables were described as the number (percentage).

Eight patients with pES and 1 with isolated autoimmune hemolytic anemia (AIHA) were treated with abatacept within the context of CTLA-4 (n = 5), LRBA (n = 3), and DEF6 (n = 1) variants. The median time from AIC initial diagnosis to molecular identification of the CTLA-4 defect was 6.9 years (range, 0.3-20.1 years). Median time to abatcept introduction was 5.5 months (range, 0.6-19.2 months) after the molecular diagnosis. The median follow-up was 4 years (range, 1.6-5.2 years) after abatacept initiation (Table 1).

At the time of abatacept initiation, the median age was 18.5 years (range, 2.3-26.6 years). Six patients had immune thrombocytopenic purpura (ITP) and AIHA in complete remission (CR), 2 had ITP in partial remission (PR) or no remission (NR; patients 2 and 3), 1 had AIHA in NR (patient 9); and 8 of 9 patients had a median number of 3 active IMs (range, 1-5), mainly pulmonary and gastrointestinal. All patients had long-lasting hypogammaglobulinemia and received antibioprophylaxis and immunoglobulin replacement. All patients had received immunoglobulin or steroids for ITP or AIHA and a median number of 4 (range, 0-8) previous second-line treatments (rituximab, n = 7; splenectomy, n = 3). In 3 patients, hydroxychloroquine was started concomitantly; in 5 patients, ongoing treatment was steroids (n = 3), Rapamune (sirolimus; n = 2), everolimus (n = 1), or romiplostim (n = 1).

Initial doses for abatacept were 10 (n = 4) or 20 mg/kg (n = 5). In patients receiving 10 mg/kg, doses were increased to 20 mg/kg because of gastrointestinal flare-up at month 2 (patient 8), progressive AIHA at month 4 (patient 9), progressive lung involvement at month 8 (patient 1), and ITP relapse at month 39 (patient 2). Initially, IV injections were repeated every 2 weeks and then progressively spaced out. However, injections had to be maintained more frequently in 3 patients: ITP flare-up (patient 2) and persistent pulmonary (patient 1) or gastrointestinal disease (patient 7). The median duration of abatacept was 3.2 years (range, 1.1-4.9 years).

Regarding AIC, CR was obtained in patient 2, and sustained PR was obtained in patients 3 and 9. Moreover, no relapse was observed despite cessation or decrease in ongoing treatments in all 6 patients already in CR. Regarding active IMs, significant improvement was seen in 7 of 8 patients, particularly in lung disease. The benefit was enhanced when abatacept was infused every 2 weeks. Abatacept allowed steroid, Rapamune, or everolimus discontinuation (patients 2, 7, 8, and 9) and Rapamune or romiplostim sparing (patients 3 and 9). In patient 6, abatacept was replaced by vedolizumab after 3.1 years because of refractory enteropathy.

No severe adverse effects occurred. Three patients had recurrent mouth ulcers, which were attenuated by a reduction in abatacept dose and/or by increased immunoglobulin replacement therapy.

At the last follow-up, all patients were alive, and 8 patients were on continuous treatment. None of the 9 patients had presented with new ITP or AIHA flare-ups. Patient 9 was a candidate for hematopoietic stem cell transplantation (HSCT) because of initial extreme severity and persistent hemolysis. No patient, excluding patient 6, had been treated with additional immunosuppressive therapy or splenectomy.

AIC in patients with insufficient CTLA-4 often had a severe, chronic, and relapsing course.¹³ We have herein reported a national prospective observational analysis of 9 children and young adults with refractory AIC and IM despite extensive treatments. The response to abatacept was excellent and so was the long-term benefit-risk ratio with treatment duration up to 4.9 years. A significant sparing of associated immunosuppressive therapies was also noted.

Similar results were reported in a first case report in 2016.¹ In 2020, 22 Turkish children with deficient LRBA received abatacept for a shorter median duration of 12.5 months, and regarding the IMs, the best response was observed in AIC.¹⁴ 

In our study, as previously reported in the context of rheumatoid arthritis, drug efficacy appears to be dose dependent, with a better response rate with 1-week or 2-week interval in comparison with 4 weeks. An initial dose of 20 mg/kg per 15 days until complete response of all target organs appears to be well tolerated and facilitates progressive spacing of infusions as proposed in juvenile arthritis.¹⁵ The tolerance was excellent for treatment duration up to 5 years, assuming that optimal prevention of recurrent infections and nonmalignant lymphoproliferation is provided by immunoglobulin replacement and antibioprophylaxis. In the field of pediatric rheumatic diseases, with up to 5 years of follow-up, the long-term safety profile of abatacept is favorable, especially for the risk of malignancies.^16-19 The risk of relapse on abatacept discontinuation is not known. Thus far, HSCT is the only long-term cure for CTLA-4 insufficiency and LRBA deficiency,^20,21 and abatacept before HSCT may improve organ function and lead to superior outcomes.

A multidisciplinary approach is warranted in patients with AIC associated with hypogammaglobulinemia and/or severe multiorgan manifestations. It is essential to perform genetic studies looking for an underlying PID. Our results clearly demonstrate that patients with different CTLA-4 pathway defects, even young children, may benefit from early treatment with off-label new targeted treatment options to avoid heavy nonspecific immunosuppression and the development of severe multiorgan manifestations. National and international collaborative cohorts or registries will allow validation of therapeutic guidelines and standardization of evaluation criteria for the various organ manifestations in those diseases. Two ongoing prospective trials may answer questions raised by our work on the benefit-risk ratio of abatacept and on its specific impact in AIC (clinicaltrials.gov: #NCT03733067; ABACHAI: EudraCT 2019-000972-40).