The prevalence, clinical relevance and origin of autoantibodies in patients with Common variable immunodeﬁciency on regular immunoglobulin replacement therapy

Background: Common variable immunodeﬁciency (CVID) is an inborn error of immunity characterized by disturbed immunoglobulin production. Despite of the terrain with severe antibody deﬁciency, autoantibody-mediated autoimmune phenomena belong to the most frequent autoimmune manifestation. However, many unresolved issues such as prevalence, clinical relevance and origin of autoantibodies detected in CVID patients receiving immunoglobulin replacement therapy (IRT) make the diagnostics of autoimmune complications diﬃcult. Methods: A prospective observational study evaluating the spectrum of 38 diﬀerent autoantibodies in 38 CVID patients receiving IRT, and in the immunoglobulin solutions used for IRT. Results: The study reveals a high prevalence of anti-GAD (55.3%) and anti-TPO (68.4%) autoantibodes in the cohort of 38 CVID patients on regular IRT. However, the titers of anti-GAD (3.22 vs. 22 kU/L, p[?]0.0001) and anti-TPO (109.7 vs. 713 kU/L, p[?]0.0001) were signiﬁcantly lower compared to the newly diagnosed T1D and AIT patients. Moreover, none of the CVID patients with detectable antibodies manifested with T1D and only three patients became suspected of having AIT. A high quantity of anti-GAD (3.24-24.48 kU/L) and anti-TPO (123.6-156.55 kU/L) autoantibodies was found in immunoglobulin solutions for IRT. Conclusions: The study ﬁnds a very high prevalence of anti-GAD and anti-TPO autoantibodies in CVID patients receiving regular IRT. Nevertheless, the presence of anti-GAD and anti-TPO is not associated with the manifestation of the respective autoimmune disease. As the high titers of both anti-GAD and anti-TPO were also found in the therapeutics used for IRT, we suggest that the therapeutic immunoglobulins are the source of this false positivity.


Introduction
Common variable immunodeficiency disorder (CVID) is one of the most frequent inborn error of immunity with estimated prevalence 0.6 -3.8 / 100 000 in European coutries (1).It is characterized by decreased immunoglobulin production, impaired specific antibody response and higher susceptibility to infections along with immune system dysregulation and higher prevalence of non-infectious complications, including a broad spectrum of autoimmune, lymphoproliferative and granulomatous manifestations (2)(3)(4), which significantly contribute to the morbidity and mortality of CVID patients (5,6).A recent meta-analysis showed the pooled prevalence of autoimmunity is 29.8% in CVID (ranging from 13.6% to 54.1%) (7).Despite the markedly impaired antibody production, diseases hallmarked by the presence of autoantibodies, such as Autoimmune hemolytic anemia (AIHA) or Immune thrombocytopenic purpura (ITP), are amongst the most commonly diagnosed autoimmune complications in CVID patients, found in up to 38.9% of patients (7,8).An increased prevalence of other autoimmune diseases associated with the tissue specific autoantibodies in comparison to general population was also reported, such as Autoimmune thyroiditis (AIT) or Type 1 diabetes (T1D) revealed in 0.8-8.7%, and 0.8-7.1% of CVID patients respectively (7,9,10).The mainstay of CVID management is a regular, long-term immunoglobulin replacement therapy (IRT).Importantly, the immunoglobulin solutions used for IRT were shown to contain various specific antibodies and may even be responsible for IRT associated adverse events such as self-limiting acute hemolysis triggered by pasivelly transmitted antierythrocyte alloantibodies (11).Therefore, we initiated a prospective observational trial to determine the prevalence, clinical significance and the origin of selected spectrum of autoantibodies in the sera of CVID patients.

Methods
This study was designed as a prospective observational trial and was approved by the Ethical Committee of Motol University Hospital.Only patients with signed informed consent and fulfilling inclusion criteria (1.ESID / ICON diagnostic criteria for CVID (3,4), at least 3-month interval of follow up and exposition to the regular immunoglobulin replacement therapy) and exclusion criteria (previous exposition to the immunosupressive/corticosteroid therapy and/or the diagnosis of AIT or T1D) were enrolled and followed for up to 2 years.
Apart from the spectrum of autoantibodies the parameters of glucose and inzulin metabolism -fasting serum concentration of C-peptide (C-peptide Kit, Roche, Mannheim, Germany) measured by CLIA (Cobas e601, Roche, Mannheim, Germany) and glycosylated hemoglobin A1c (g-Hgb; Capillarys Hb A1c Kit, Sebia, France) using capillary electrophoresis(Capillarys 2 Analyzer, Sebia, France); thyroid gland function -free thyroxine (fT4; ADVIA Centaur FT4 assay), thyroid stimulating hormone (TSH; ADVIA Centaur FT4 assay) assessed by CLIA on ADVIA Centaur XPT Systems (Siemens, Tarrytown, NY). and thyroid gland ultrasonography (USG; Toshiba Nemio MX, Tokyo, Japan) were performed at screening and then in a year-long intervals and at the end of the study.Clinical follow-up visits were conducted in 3-month intervals with IgG serum levels assessment using nephelometry (IMMAGE 800, Beckman Coulter, Indianapolis, IN, USA) and Human Serum IgG Kit (Beckman Coulter, Indianapolis, IN, USA).
The selected laboratory parameters were also compared to a cohort of 40 newly diagnosed T1D and 50 AIT patients.
Mean values and standard deviations (SD) were calculated.Two-Sample T test was used for unpaired parametric, Paired T test for paired parametric, and Kruskal-Wallis for multiple non-parametric data set.The differences were statistically significant when p value was [?]0.05).Statistical analysis was perfomed in Minitab, version 17.1 (Minitab Inc., State College, PA, USA).

Baseline characteristics
Thirty-eight patients were enrolled, 22 females and 16 males, two patients with previously diagnosed Type 1 diabetes and 3 patients with Autoimmune thyroiditis were exluded at screening.The mean age at study initiation was 37.5 years (+/-13.31SD, range: 17-75), the mean time of disease duration 10 years (+/-7.59,range: 0.25-28).Twelve patients were treated with IVIG and 26 patients with SCIG, the mean dose 300mg/kg/month (+/-59.6.range: 200-420).The dose of IRT remained unchanged during the duration of the entire study.The mean time of the participation in study was 2 years (+/-0.64,range: 0.25-2).There were no statistically significant differences in baseline characteristics between patient groups with negative and positive values of anti-GAD (>0.9 kU/mL) and anti-TPO (>60 kU/mL) autoantibodies respectively (Table 1).

The detection of autoantibodies in immunoglobulin therapeutics
The spectrum of autoantibodies including ANA, anti-dsDNA, anti-EMA, RF, ACLA, anti-F-actin, anti-SLA, anti-LKM1, anti-LC1, anti-AMAM2, anti-MPO, anti-PR3, anti-Sm, anti-RNP, anti-SS-A, anti-SS-B, anti-Scl70, anti-Jo-1, anti-IAA, anti-IA2 and anti-TG were not detected or were found to be below the lower limit of reference value (with the exception of anti-TG in SCIG-16.5 solution).In contrast, the anti-GAD and anti-TPO were detected in very high levels in all the teste IRT.The titers of anti-GAD were fluctuating from 3.24 kU/L (+/-1.51,range: 0.9-4.74) in 10% IVIG-I to 24 2).Only the differences in the content of anti-GAD (p <0.0001) between particular therapeutics were statistically significant (Supplementary Figure 4 and 5).

Discussion
CVID represents one of the most common primary antibody deficiency which is associated with a broad spectrum of non-infectious complications including various autoantibody-mediated autoimmune diseases.Several studies reported the presence of various antibodies directed against "self-antigens", such as feritin, thyreoglobulin or DNA in the therapeutics used for IRT (12).Antierythrocyte alloantibodies or anti-Ro (SS-A) were even associated with clinical manifestation -self-limiting hemolysis (11) and sicca syndrome (13).Thefore, we initiated the prospective observational study focusing on the prevalence, clinical significance and the origin of selected spectrum of autoantibodies in a cohort of 38 CVID patients treated with regular IRT.
Anti-GAD and anti-TPO were the most prevalent autoantibodies found in the majority of CVID patients on regular IRT (68.4% and 55.3%, respectively).All other investigated autoantibodies including ANA, anti-dsDNA, anti-EMA, RF, ACLA, anti-F-actin, anti-SLA, anti-LKM1, anti-LC1, anti-AMAM2, anti-MPO, anti-PR3, anti-Sm, anti-RNP, anti-SS-A, anti-SS-B, anti-Scl70, anti-Jo-1, anti-IAA, anti-IA2 and anti-TG were negative in all patients, with exception of anti-TG, anti-IAA and anti-IA2 found in patients with previously diagnosed T1D and AIT at the screening visit, who were excluded based on exclusion criteria.Interestingly, both excluded CVID patients with T1D were anti-GAD negative.However, the titers of anti-GAD observed in our cohort of CVID patients were significantly lower in comparison to newly diagnosed T1D patients.The cohort of CVID patients was prospectively followed to assess the T1D development.This included regular measurements of fasting C-peptide as a marker of inzulin production and g-Hgb reflecting long-term glycemia.No differences were found in serum concentration of C-peptide and g-Hgb between anti-GAD negative and positive CVID patients, as well as between the baseline and end-of-study visits in anti-GAD positive patients.On the other hand, the serum levels of C-peptide were significantly higher and the levels of g-Hgb significantly lower in comparison to newly diagnosed T1D patients suggesting an undisturbed inzulin production and glucose metabolism.
Similarly, the serum titers of anti-TPO autoantibodies in CVID patients were also significantly lower than in newly diagnosed AIT patients.However, no differences were found in the serum levels of TSH and fT4 between anti-TPO positive, anti-TPO negative CVID patients and AIT patients.Both paramters were within referential limits in all groups.Therefore, the thyroid gland USG was performed to evaluate the presence of predictive signs of subclinical AIT (14).Based on the USG, the diagnosis of AIT was considered in three out of 26 anti-TPO positive patients and in one patient without AIT specific autoantibodies.
The same spectrum of autoantibodies was also assessed in IRT therapeutics.Surprisingly, all solutions contained high amount of anti-GAD and anti-TPO autoantibodies.Other investigated autoantibodies were uniformly negative.The possible explanation for this phenomenon may be found in population-based studies investigating the prevalence of anti-GAD and anti-TPO.Anti-GAD autoantibodies were present in 0.9 -1.7% of healthy donors, the highest prevalence 3.2% was observed in age group 30-34 years (15).Even higher prevalence was reported for anti-TPO antibodies -up to 31.7% of healthy donors (16,17).Based on our findings and the previously published reports, we therefore suggest that these autoantibodies may be passively transfered into blood circulation via the immunoglobulin solutions.Our findings are also consistent with previous observations that T1D and AIT are predominantly T-cell-mediated diseases (18)(19)(20)(21).Correspondingly, both T1D and AIT may develop in the abscence of specific autoantibodies.The autoantibody negative cases comprise 3.5-19% from all T1D patiens (22,23), moreover, several seronegative T1D patients were also desribed amongst the CVID patients (10).Autoantibody negative AIT represents approximately 5% of all patients with AIT (24).While the tissue specific autoantibodies may not constitute the principal pathophysiological mechanisms in T1D and AIT, they represent a commonly used diagnostic and prognostic marker.The physicians should therefore be sensitized to the fact, that in CVID patients on regular immunoglobulin subtitution therapy neither serum anti-GAD nor anti-TPO are suitable tools for the screening or diagnosis of T1D or AIT.
The authors are aware, that due to the limitations of this pilot study, the results must be interpreted with caution.The strength of the study is particularly in its prospective design and the utilization of routine and certified laboratory methods.On the other, the limited number of included patients and the relatively short trial follow up time are its main limitations.Further multicentric and longer follow up studies are warranted to confirm these findings.

Conclusions
To our best knowledge, this is the first report investigating the content of broad spectrum of autoantibodies in therapeutics used for IRT and their clinical relevance.We identified high quantities of anti-GAD and anti-TPO autoantibodies in all investigated therapeutics that may be passively transfered to the patients' blood circulation.
Despite the fact that IRT may lead to autoantibody transfer, we found no evidence that this mechanism would contribute to the clinical manifestation of the autoimmune diseases, compromising the safety of IRT.However, it might interfere with the disease diagnosis.Based on our results, we recommend that anti-GAD and anti-TPO should not be used for the screening or diagnosis of T1D and AIT in CVID patients on regular immunoglobulin subtitution therapy.Instead, the measurement of glycemia, C-peptide, gHb and thyroid ultrasound, respectivelly, may be the screening methods of choice.

Akcnowledgement
The study was supported by the grant of the Czech Health Research Council nr.NV18-05-00162 and the Technology Agency of the Czech Republic nr.TJ04000443.

Authors Contribution
TM (main author) made substantial contributions to conception and design of the study, analysis and interpretation of data, drafted the manusript KK (co-author) made substantial contributions to conception and design of the study, acquisition of data, or analysis and interpretation of data PL (co-author) made substantial contributions to conception and design of the study, acquisition of data, or analysis and interpretation of data JS (co-author) made substantial contributions to conception and design of the study, acquisition of data, analysis and interpretation of data, drafted the article MB (co-author) made substantial contributions to conception and design of the study, acquisition of data, analysis and interpretation of data, drafted the article Figures Figure 1: The serum levels (mean, 95% confidence intervals and individual symbols) of anti-glutamic acid decarboxylase autoantibodies (Anti-GAD) in anti-GAD positive (> 0.90 kU/L) Common variable immunodeficiency patients (CVID+) compared to newly diagnosed patients with Type 1 diabetes (T1D).Supplementary Figure 4: The concentration of of anti-glutamic acid decarboxylase (anti-GAD) autoantibodies (mean, 95% confidence intervals and individual symbols) in different solutions for immunoglobulin replacement therapy (10% solution for intravenous use IVIG-I and IVIG II, 16.5% and 20% solutions for subcutaneous use SCIG-16.5 and SCIG-20), statistically significant differences (p [?]0.0001), upper reference limit (0.9 kU/L) indicated by bold gridline.

Figure 2 :
Figure 2: The serum concentration (mean, 95% confidence intervals and individual symbols) of C-peptide (C-p) in anti-GAD negative (CVID-) and positive Common variable immunodeficiency patients (CVID+) and at baseline (BSL), end-of-study (EOS) visits compared to newly diagnosed patients with Type 1 diabetes (T1D) are displayed, upper (1470 pmol/L) and lower (370 pmol/L) reference limits indicated by bold gridlines.