Autoimmune and autoinflammatory diseases
Home · Research areas

Immunomodulatory treatments

Low-dose interleukin-2, azacitidine, biologics and signalling pathway inhibitors: treatments that seek to correct a disordered immune response rather than to switch it off. This page brings together what the work of Arsène Mekinian has measured on these strategies, and what it has not established.

What is an immunomodulatory treatment?

The immune system has cells whose task is to restrain its own reactions, the regulatory T lymphocytes. When this brake is insufficient, the immune response can turn against the body, or against the pregnancy. An immunomodulatory treatment does not seek to suppress this response as a conventional immunosuppressant would: it seeks to bring it back into balance. Interleukin-2 is a natural messenger of the immune system. Given at a very low dose, it stimulates mainly the regulatory T lymphocytes. It must be distinguished from high-dose interleukin-2, used in oncology, which belongs to an entirely different use.

A second and different idea concerns certain inflammatory diseases of adults: the inflammation there arises from a clone of bone marrow cells carrying an acquired mutation. Acting on the inflammation alone is then not enough, and the strategy consists in targeting the clone itself, with a hypomethylating agent such as azacitidine. A third family, the biologics and the signalling pathway inhibitors, blocks one precise molecule or pathway. None of these approaches holds good in principle. Each is measured in patients, on an endpoint defined in advance, and that is what the works described below do.

His work on this subject

The TRANSREG trial assessed low-dose interleukin-2 in eleven autoimmune diseases brought together in a single open-label trial, which was the question asked: can the same dose act on the same mechanism in different diseases? Forty-six patients with a mild to moderate form of rheumatoid arthritis, ankylosing spondylitis, systemic lupus erythematosus, psoriasis, Behçet's disease, granulomatosis with polyangiitis, Takayasu arteritis, Crohn's disease, ulcerative colitis, autoimmune hepatitis or sclerosing cholangitis received 1 million units per day for five days, then one injection every two weeks for six months. The treatment was well tolerated whatever the disease and the concomitant treatments, and a specific expansion and activation of regulatory T lymphocytes was observed in all patients, without activation of effector T lymphocytes. The authors write that an indication of potential clinical efficacy was observed and conclude that this result justifies launching phase II efficacy trials. Arsène Mekinian signs for the department of internal medicine of Saint-Antoine Hospital, 17th of 27 named authors (Annals of the Rheumatic Diseases, 2019, trial NCT01988506). The same design was then broadened: an open-label phase 2a basket trial treated 81 patients with one of 13 autoimmune diseases, with the primary endpoint being the change in regulatory T lymphocytes at day 8. The expansion was found in all 13 diseases, without activation of effector lymphocytes, and the disease-specific clinical scores improved in five of the six cohorts with at least six patients, namely ankylosing spondylitis, lupus, Behçet's disease, Sjögren's syndrome and systemic sclerosis. Urticaria was the only serious treatment-related adverse event. He is 13th of 21 named authors (Journal of Autoimmunity, 2024).

It is in systemic sclerosis that he carries the work as last author and corresponding author. Nine patients without severe organ involvement, included within TRANSREG, received 1 million international units per day for five days, then one injection every two weeks for six months. The primary endpoint was met at day 8: the frequency of regulatory T lymphocytes among CD4-positive T lymphocytes increased by a factor of 1.8 plus or minus 0.5 (p = 0.0015), without significant change in effector T lymphocytes or in B lymphocytes. The modified Rodnan skin score and the Valentini score remained broadly stable at six months, as did the activity and severity measures, quality of life and pulmonary function tests. The treatment was well tolerated, with no serious treatment-related adverse event. The authors conclude that this result opens the way to properly sized phase II trials, in other words that clinical efficacy remains to be demonstrated. He is 11th and last of 11 named authors, and corresponding author (RMD Open, 2024).

Two publications describe the use of low-dose interleukin-2 outside the framework of these trials. The first reports three patients with a myelodysplastic syndrome and an associated dysimmune manifestation, respectively polymyalgia rheumatica, relapsing polychondritis with Sweet's syndrome, and cutaneous and joint vasculitis. All three were steroid-dependent and refractory to biologics or to azacitidine. They received 1 to 1.5 million units per day for five days, then every two weeks. The treatment allowed clinical improvement and steroid sparing in two of the three patients, without serious adverse event and without disease progression, and the authors conclude that a controlled study is needed. He is 8th of 9 named authors, the article also carrying the signature of the MINHEMON network, which does not count towards the number of authors (Rheumatology, Oxford, 2021). The second is a 2022 letter devoted to a cohort of off-label use of low-dose interleukin-2 in systemic autoimmune diseases, of which he is 6th and last named author and corresponding author. The MEDLINE record of this letter carries no abstract: no figure from it is reproduced here (Clinical and Experimental Rheumatology, 2022).

In reproductive immunology, the question has been raised in his work for a long time. In 2016 he is 1st of 9 named authors of a review with meta-analysis devoted to the place of immunomodulation in recurrent miscarriage and unexplained recurrent implantation failure. The figures it brings together come from the literature analysed and not from his own patients: a modest benefit of progesterone on live birth, with an odds ratio of 1.38 (95% confidence interval: 1.07 to 1.77) and substantial heterogeneity between studies (p = 0.01, I2 = 78%). The conclusion calls for the efficacy of these treatments to be demonstrated and for the subgroups of patients concerned to be defined (American Journal of Reproductive Immunology, 2016). The FaCIL-2 trial (NCT03970954) took this question to low-dose interleukin-2. It is a phase 1 and 2, open-label, single-group trial, sponsored by AP-HP with the company Iltoo Pharma as collaborator, conducted at three AP-HP sites, Saint-Antoine, Tenon and Trousseau, in women aged 18 to 40 with at least five consecutive early miscarriages before 14 weeks of amenorrhoea, unexplained after investigation. The principal investigator is Prof. David Klatzmann. Arsène Mekinian is study director there, the registry's "Study Director" entry, and Saint-Antoine site lead, alongside Prof. Gilles Kayem for Trousseau. The registry shows 18 participants enrolled, a study start on 5 January 2021, a study end on 20 March 2024, and no results posted.

The results of this trial are available only as a preprint, posted on medRxiv on 16 June 2025, of which Arsène Mekinian is the 1st of eight authors. This preprint is not peer-reviewed: no independent reviewer has checked its data or its analyses, it may still be modified, and it is not to be cited as an established scientific publication. It describes 15 women treated with 3 million units per day for five days per cycle, from the tenth day after the start of menstruation, for at most five cycles. The primary endpoint was met, with a mean 2.0-fold increase in circulating regulatory T lymphocytes at day 8 (p below 0.001). Eight pregnancies are reported, four of them ongoing at 12 weeks, leading to three live births and one late miscarriage at 20 weeks. Nine further patients, treated outside the trial on compassionate grounds and without immunological follow-up, had five pregnancies and two live births. The medRxiv page has since flagged a one-page text published in June 2026 in a supplement of La Revue de médecine interne (volume 47, supplement S1, page A49, five authors, Arsène Mekinian first): this text is not indexed in PubMed and does not amount to full publication of the results in a peer-reviewed journal.

On the haematological side, the strategy changes target: the aim is to act on the marrow clone. He is 1st of 36 named authors, and corresponding author, of a prospective phase II trial of azacitidine in steroid-dependent or refractory systemic autoimmune and inflammatory manifestations and in VEXAS syndrome associated with myelodysplastic syndromes and chronic myelomonocytic leukaemia, published in 2022 as a letter in Leukemia. The MEDLINE record of this letter carries no abstract, so no figure from it is reproduced here. Two registry studies complete this picture. In 2022, from a French national registry of 116 patients with VEXAS syndrome, azacitidine was assessed in 11 patients with an associated myelodysplastic syndrome, with a clinical response in five of them, that is 46%; he is 17th and last of 17 named authors (British Journal of Haematology, 2022). In 2025, the multicentre retrospective study of the FRENVEX group covered 88 patients with genetically confirmed VEXAS syndrome, 80% of whom met the 2022 WHO criteria for a myelodysplastic syndrome. The inflammatory response was 41% at six months and 54% at twelve months. A molecular response, defined as a fall of at least 25% in the variant allele frequency of UBA1, was observed in 65% of patients, and red blood cell transfusion independence in 65% as well. Infections, in 34%, and cytopenias, in 36%, were frequent, mainly during the first three cycles. The authors conclude that these results justify larger prospective trials. He is 53rd of 54 named authors (Blood, 2025).

A strategy is judged only on a clinical endpoint set in advance, and his work shows that the result is not always the one expected. In the TOCITAKA trial, prospective, multicentre and open-label, 13 patients with Takayasu arteritis who had never been treated received corticosteroid therapy at 0.7 mg per kilogram per day and seven monthly infusions of tocilizumab at 8 mg per kilogram. Six of the 13 patients, that is 54%, met the primary endpoint, namely withdrawal of corticosteroids after seven infusions, with a significant fall in disease activity at six months. Among the 11 patients in remission at six months, five, that is 45%, relapsed within the 12 months following withdrawal of tocilizumab, which leads the authors to conclude that maintenance treatment is needed. He is 1st of 21 named authors and corresponding author, the article also carrying the signature of the French Takayasu network (Arthritis Research and Therapy, 2020, trial NCT02101333). Conversely, the double-blind placebo-controlled randomised trial that assessed interleukin-6 receptor blockade in primary Sjögren's syndrome, in 110 patients divided between tocilizumab and placebo, showed no improvement: at week 24, the primary endpoint was met in 52.7% of patients in the tocilizumab group, that is 29 of 55, and in 63.6% of the placebo group, that is 35 of 55. He is 11th of 39 named authors (Annals of the Rheumatic Diseases, 2021, trial NCT01782235). Finally, an international retrospective analysis of 30 patients with VEXAS syndrome treated with various JAK inhibitors concluded in favour of ruxolitinib, with clinical remissions and a reduction in corticosteroid therapy in most of the patients treated with this molecule; he is 6th of 25 named authors (Blood, 2022).

Where care is organised

These treatments are not routine treatments. They belong either to registered clinical trials, with inclusion criteria published at the same time as the trial, or to collective decisions taken disease by disease and situation by situation.

The works cited above were conducted from the department of internal medicine of Saint-Antoine Hospital (AP-HP, Sorbonne Université), attached to the reference centre for autoinflammatory diseases and inflammatory amyloidosis. For the whole interleukin-2 strand, they involve the biotherapy department and the clinical investigation centre in biotherapy of the Pitié-Salpêtrière hospital, which carries out the immunological follow-up of patients, as well as the clinical research unit which carries out the statistical analysis. The FaCIL-2 trial brought together three AP-HP hospitals, Saint-Antoine, Tenon and Trousseau, with AP-HP as sponsor.

The work on dysimmune manifestations associated with myeloid haematological disorders and on VEXAS syndrome runs through national multicentre networks, named in the publications themselves: the MINHEMON network, the FRENVEX group for VEXAS syndrome, the Groupe francophone des myélodysplasies, and the French Takayasu network for large-vessel vasculitis. It is these networks that make it possible, in rare diseases, to assemble numbers large enough to measure something other than an impression.

What is not known

The clinical efficacy of low-dose interleukin-2 is not demonstrated. None of the published trials in which he took part has a control group: TRANSREG and the 2024 basket trial are open-label trials whose primary endpoint is immunological, the change in regulatory T lymphocytes, and not clinical. The conclusion of the 2019 article is itself that phase II efficacy trials remain to be launched. In systemic sclerosis, nine patients were treated and the scores remained stable over six months: stability observed without a comparison group is not attributed to the treatment.

In recurrent early pregnancy loss, the point is sharper still. The FaCIL-2 trial is open-label, single-group, and its results are not peer-reviewed. The registry indicates 18 participants enrolled, the preprint analyses 15: this discrepancy is not explained by the public documents. The number of pregnancies and live births reported concerns a few women, without a comparator, and allows no estimate of a success rate. The nine patients treated outside the trial on compassionate grounds do not constitute a practice and are not trial data. The preprint itself concludes that a controlled trial is needed, with prolonged treatment at the start of pregnancy. In other words, the effect of low-dose interleukin-2 on live birth remains an open question, and this approach has no validated indication in this situation.

The data outside trials rest on very small numbers, three patients for the 2021 report, and on a 2022 letter whose abstract is not available in the MEDLINE record. For azacitidine, the 2025 FRENVEX study is retrospective and without a comparator, and its authors themselves call for larger prospective trials; the numerical results of the prospective phase II trial published in 2022 do not appear in its MEDLINE record and are therefore not reproduced here. The study of JAK inhibitors in VEXAS syndrome is retrospective and covers 30 patients: the choice of drug there lay with the doctor, which introduces a selection that the analysis does not correct.

Finally, a plausible mechanism guarantees nothing. The randomised placebo-controlled trial of interleukin-6 receptor blockade in Sjögren's syndrome, in which he took part, showed no improvement at week 24, with a numerically more favourable result in the placebo group. In Takayasu arteritis, nearly half of the patients in remission relapsed after withdrawal of tocilizumab. The question that remains unanswered, for all these strategies, is that of patient selection: no marker published in this work makes it possible today to predict who will respond, or for how long. This work is moreover collective, signed by dozens of authors within multicentre networks; the position of Arsène Mekinian, given for each reference below, states the part he takes in it.

References

Published work he has taken part in, on this subject.

General information only. This page does not replace individual medical advice. Updated: 10 September 2026.