Studies on biological treatment and biomarkers in systemic lupus erythematosus : longterm effects, risk factors and predictors of response

Abstract

Systemic Lupus Erythematosus (SLE) is an autoimmune disease in which several immune mechanisms are involved in a complex interplay. B-cells are considered to be a major player in the pathophysiology of the disease, and the advent of B-cell depletion therapy (BCDT), has brought much promise. Yet, after twenty years of use of the depleting agent Rituximab (RTX), BCDT still has to find its precise location and role in SLE therapy. Among SLE clinical manifestations, lupus nephritis (LN) carries a significant burden in terms of morbidity, and requires optimised approaches. Identifying clinical and biological biomarkers, may contribute to improving the clinical management of the disease and its major organ involvement.

The aims of this thesis were the following: 1) to contribute to a definition of RTX’s role in SLE treatment, by studying its deep immunological effects, immunogenicity and side effects, aiming at identifying possible biomarkers of efficacy and safety which might be implemented in daily practice; 2) to identify possible non-invasive biomarkers of specific organ involvement, such as LN, which may also in the near future be implemented in daily practice for identifying LN, and to monitor disease activity and response to treatment.

In Paper I, deep immunological effects of RTX on recently defined B- and T- cell subsets were explored, through a multicolour flow cytometry approach. In particular, we investigated whether RTX affects the age-associated B-cells (ABCs) which are plasma cells precursors; and the T-follicular and T-peripheral helper T-cells (TFH, TPH). We showed that transient reduction of the frequencies of the B-cell phenotype age-associated B-cells (ABCs) is induced by the treatment during the early phases of B-cell depletion. This corresponded to a reduction within the DN2 compartment, which contains the ABCs cells. By contrast, no early significant changes of the TFH and TPH followed the administration of RTX. A reduction of a subset of CD4+ with high expression of the marker PD-1 was shown at later follow-up. Examining the behaviour of these cell subsets in patients who developed antibodies against the drug, we showed that immunogenicity was associated to lower frequencies of double negative (DN) B-cells and to a wider expansion of plasma blasts at early follow up.

In Paper II, we investigated the occurrence of anti-drug antibodies (ADAs) towards RTX, in a cohort of 66 SLE patients after their first course of treatment, in comparison with 22 first-ever treated ANCA-associated vasculitis (AAV) patients. After the first course of RTX almost 38% of SLE patients developed ADA, while in AAV no ADA were shown. The SLE patients developing ADA were younger, had a longer disease duration, and a more active disease at the time of treatment initiation. They were mostly treated for LN and had a serologically active profile. Despite an overall reduction of disease activity upon RTX treatment, the presence of ADA in SLE was associated with higher counts of B-cells in the peripheral blood at the 6 month follow-up, and with higher residual disease activity. At retreatment, ADA-positive SLE patients experienced more frequent infusion reactions, both of immediate and late-onset occurrence.

In Paper III, we investigated the occurrence of late onset neutropenia (LON) in 107 patients with SLE treated with RTX. We found a rate of occurrence as high as 30%. In most cases the laboratory finding was discovered in routine controls and did not cause complications. However, around 40% of the patients experienced clinical consequences with occurrence of neutropenic fever and infections, some requiring intensive care. The clinical characteristic associated with the risk of LON was high disease activity before treatment. At a biological level, the appearance of LON was associated with the increase of BAFF which follows the disappearance of B-cells in the peripheral blood.

In Paper IV, we explored the use of the soluble protein Galectin-3 Binding Protein (Gal- 3BP), as urinary biomarker in LN, comparing the urinary concentration of this protein in 86 patients with active LN, 63 patients with active SLE without LN, 73 inactive SLE patients, and 48 matched population based controls. Samples were tested separately for the concentrations of other biomarkers: Galectin-3 (Gal-3), neutrophil gelatinase associated lipocalin (NGAL), osteopontin (OPN) and kidney injury molecule-1 (KIM-1). Each of the biomarkers was assessed as absolute concentration and as concentration normalized for the urine-creatinine (adjusted concentration). We found that the levels of both non-adjusted and adjusted u-Gal-3BP were significantly higher in the urine of patients with active LN as compared to all the other groups. Higher levels of adjusted u-Gal-3BP were found in proliferative and membranous forms of LN as compared to mesangial forms. We also found a moderate correlation between the adjusted u-Gal-3BP and the histological activity index as evaluated in kidney biopsies. This correlation was stronger when we considered patients not receiving immunosuppressive treatments at the time of kidney biopsy. Current treatment with oral corticosteroids was associated with lower urinary levels of Gal-3BP in active LN patients. In patients with proliferative LN this association was found also regarding ongoing treatment with antimalarials. In a subset of ten patients with active LN, significant reduction of Gal-3BP levels was observed at repeated analysis after treatment.

In Paper V, we explored the presence in the urine of LN patients (n=13) of extracellular vesicles (EVs) carrying an array of molecules of interest, all with a putative role in inflammation. We found that all the tested EVs were detectable in the urine, although at lower concentrations as compared to blood. We preliminary adopted a cut-off of EVs concentration of 50x10^6/L and evaluated the concentration of EVs carrying the cargo molecules of interest above the above mentioned cut-off. We considered the expression of EVs with respect to the histological activity. No correlation was found, but the concentration of EVs carrying the split complement molecule C5a, was found to be significantly higher in patients classified as active at renal biopsy as compared to those classified as inactive. We then examined the expression of urinary EVs with respect to having a predominant proliferative histologic pattern with respect to non-proliferative pattern. In proliferative LN, urinary EVs expressing C3a, C4, C4d, C5a, Mitochondrial antigens, Lactadherin, NGAL and TWEAK were found in significantly higher concentrations as compared to patients with non-proliferative forms.

In conclusion, the thesis here presented adds some elements of knowledge on the immunological effects and consequences of the use of RTX in SLE, and on aspects of safety associated with immunogenicity. It also adds knowledge on the magnitude of LON as a safety aspect of RTX use as a therapy for SLE. Finally, it contributes to the definition of potential urinary biomarkers which might be implemented in clinical practice in the future.