Autoimmune Neutropenias: Update on Clinical and Biological Features in Children and Adults

Authors: Francesca Fioredda, Carlo Dufour, Petter Höglund, Helen A Papadaki, Jan Palmblad

Abstract

The definition of autoimmune neutropenias (AIN) has been based on the demonstration of autoantibodies directed to various epitopes on blood neutrophils. However, this definition is probably too limited and excludes neutropenias (NPs) with a negative autoantibody test but with other phenomena that indicate an underlying autoimmune process. Examples of such AINs may be complete or incomplete systemic lupus erythematosus or other autoimmune diseases where NP is common but patients may not fulfill formal diagnostic criteria for a rheumatic disease. Recently, various inherited immune-dysregulation syndromes, such as those related to variants in, for example, TACI, BAFFR, ACKR1/DARC, LRBA, CTLA 4 genes, with dysregulated B- and T-lymphocyte functions, have been associated with concomitant AINs. Cellular immune mechanisms may also play a prominent role in the development of NP, in the presence or not of autoantibodies, in cases of large granular lymphocyte syndromes of T- and NK-cell types or in chronic idiopathic NP, particularly in adults with T-cell clonal populations. The course of AIN may differ according to age, being transient and rather uncomplicated in children, and chronic with treatment requirement in adolescents and adults. This review discusses current knowledge of AINs, including diagnostic procedures, treatments, and prognosis.

Introduction

Neutropenias (NPs) are uncommon disorders in a Western population (occurring in approximately 1% of the population).^1,2^ They are more common in African and Middle East–based populations^3^ due to the gene variation in the atypical chemokine receptor-1 (ACKR1)/duffy antigen receptor of chemokines (DARC) genes, resulting in the ethnic NP (currently proposed as ACKR1/DARC-associated NP, ADAN; personal communication with the European Hematology Association [EHA] Guidelines Group on Neutropenias).

NPs may be acute (induced by drugs or infections) or chronic. The chronic NPs can be due to genetic aberrations but the majority are acquired. Among the latter, many are considered to be caused by autoimmune humoral mechanisms.

Autoimmune neutropenias (AIN) are often defined by a reduction of the absolute peripheral blood (PB) neutrophil count (ANC) below the reference values for age^4–7^ (Table 1) together with the demonstration of autoantibodies against human neutrophil antigens (HNAs) (Figure 1).^8–13^ In addition to humoral immunity, that is, autoantibodies, cellular immune mechanism are also likely to be involved in AIN, both in immune initiation and in neutrophil destruction. A classic example of an activated cellular immunity is Felty syndrome, where NP, with or without antineutrophil autoantibodies/immune complexes, coexists with longstanding rheumatoid arthritis, splenomegaly, and the presence of activated large granular lymphocytes (LGL).^14,15^ Another example is the recognition of cell-mediated inflammatory mechanisms in chronic idiopathic neutropenias (CIN) including an emerging role of the myeloid-derived suppressor cells (MDSC).^16^ Although these mechanisms are not well defined, an emerging consensus suggests that they should be considered and incorporated into algorithms for the diagnostic NP work-up, as well as for individually tailored therapies.^17–20^

This review will cover recent advances in the understanding of AIN pathogenesis, diagnostic procedures, therapeutic approaches, and outcomes in children and adults. It will not deal with alloimmune NPs. To an extent, this review is based on the Training School on AINs held under the auspices of the COST Action 18233 European Network for the Innovative Diagnosis and Treatment (Eunet-INNOCHRON) and the current work on Guidelines for NP under the sponsorship of EHA.^20^

EPIDEMIOLOGY

Prevalence of AIN

In young children with primary AIN (pAIN), the asymptomatic clinical course and the high rates of spontaneous recovery in most hampers the prevalence assessment because of underestimation of diagnosis and loss of follow-up. Historical data suggest an annual incidence of pAIN in young children 1/100,000, while data on prevalence from the Italian NP Registry on selected populations like premature babies give an estimation of 13%.^9,21^ In adults, the prevalence is even less well defined due to the scarcity of antineutrophil antibody screening in the work-up of the majority of adult NP patients and the rarity of adults reports in registries. Data on the prevalence of chronic NP are estimated to be between 0.12% and 1.4% in the general European population.^1,2,22^

The time-course and ANC variability

AINs may be acute or chronic (>3–6 months). The chronic AINs often show considerable variability in the ANC over time, with occasional month long periods of (near)normal ANC. One example is pregnancy where, resembling the physiological trend, some patients may normalize the ANC, stop previous recombinant granulocyte-colony stimulating factor (rhG-CSF) treatment (if given), while relapsing postpartum.^23^

Sometimes, this ANC variability is prominent enough to call for investigations of cyclic NP due to ELANE mutations. However, cyclic NP is characterized of a highly regular periodicity, with nadirs strictly every third week, which is not seen in AIN.^24,25^

DIAGNOSTIC CHALLENGES

Autoantibodies to neutrophils as a diagnostic test for AIN

Antineutrophil autoantibodies in PB can be detected with direct or indirect tests.^26–28^ The indirect techniques are the gold standard for the AIN diagnosis. However, the final diagnosis is highly dependent on the quality and the frequency of test repetitions. Various factors may confer false positive as well as false negative results.

Direct antineutrophil autoantibody testing

Detection of autoantibodies bound to neutrophil surface antigens, similar to the principle of the direct autoantibodies test for autoimmune hemolytic anemia, has important limitations when applied to AINs. These include the paucity of neutrophils in the neutropenic patient’s blood, passive adsorption of immunoglobulins/immune complexes to the neutrophils’ surface mainly by Fc-receptors, causing false positive reactions, and the need to perform the test on freshly drawn blood samples. These drawbacks significantly dampen the sensitivity and specificity of this assay.

Indirect autoantibody testing

Detection of autoantibodies in patient’s PB is the basis of the indirect screening. Although low ANC number is not a limiting factor in the indirect test, the method has, nonetheless, some limitations. The most relevant is that serum/plasma is incubated with isolated neutrophils from healthy donors, which may either be untyped for HNA antigens or not screened for optimal expression level of HNAs.^29,30^ This is important because the expression of the target antigens for autoantibodies are both variable and affected by the allotype of the glycoproteins in question. This paradoxical allotype specificity of autoantibodies is particularly evident for CD16, in which autoantibodies often bind much better to the HNA-1a allele than to the HNA-1b allele.^31^ Another possible pitfall of the indirect assay is the existence, albeit at low frequency, of healthy individuals who are CD16 or CD177 null, leading to false negative results.^30^ To make sure the indirect test is reliable, proper genotyping of donors in combinations of repeating the assay over the time span of the NP duration is warranted.^9,32,33^

Ideally, a combination of indirect and direct tests in reference laboratories would achieve the highest antibody detection rate^34^; however, the high cost of these procedures and the need for freshly drawn samples do not allow the routine use of this combination strategy.

The indirect test is performed by means of the granulocyte immunofluorescence technique (GIFT), ideally in combination with the granulocyte agglutination technique (GAT).^34,35^ Based on guidelines from the pediatric literature, it is recommended to repeat the test 4–6 times over 4–6 or more months, to increase sensitivity.^9,32,33^ However, these observations underline the concept that the discrimination between pAIN and idiopathic NPs, based on the detection of these autoantibodies, can be hazardous and that the 2 diagnoses may essentially represent the same disease.^32^

Finally, 2 specific aspects of the indirect antibody test need to be mentioned. One is the possibility that autoantibodies exclusively react to epitopes expressed on bone marrow (BM) neutrophil precursors, impeding neutropoiesis and causing PB NP.^36^ Such putative epitopes have not been characterized yet. The other pitfall is related to the presence of autoantibodies against HLA antigens, as a consequence of previous pregnancies and/or blood transfusions, that cross-react in the GAT and GIFT giving a false positive signal.^37^ A means to circumvent this problem is the inclusion of tests identifying the specificity of the autoantibodies, such as the monoclonal antibody immobilization of granulocyte antigen (MAIGA) test.^37^ In this test, several HNA epitopes (usually CD16, CD11a, CD11b, and CD177) (Figure 1) are captured by monoclonal antibodies, which allow the recognition of specific antibodies to these glycoproteins irrespective of the potential presence of antibodies to HLA. MAIGA, however, is a complex assay offered by specialized laboratories only.^34^

A promising new development in antibody diagnostics in NP is the Luminex-based assay LabScreen Multi, developed by One Lambda/Thermofisher. In this assay, individually colored beads are coupled to specific HNA alleles, making the identification of allele-specific antineutrophil antibodies fast and high throughput. It also contains several beads with combinations of HLA class I alleles and can, thus, also detect anti-HLA class I antibodies. Although the assay has been developed for alloantibody detection in TRALI diagnostics, it seems useful to identify antibodies to CD16 in pediatric samples.^38^ but its performance over a large number of patient samples is under validation.

MECHANISMS FOR AIN

Do autoantibodies to neutrophils cause the NP?

The mechanisms for AIN remain largely elusive. Two key questions will be discussed here. The first is where in the body neutrophils are removed, resulting in NP. The second issue is how neutrophils are removed/destructed.

Where is the site of neutrophil destruction?

Two removal sites have been the BM during neutrophil maturation and the monocyte/macrophage system, primarily of the spleen, for aged or otherwise tagged neutrophils.^39–41^

During normal maturation in the BM, many neutrophils probably undergo apoptosis, without having transmigrated to the blood. A pathological form of this process characterizes the WHIM syndrome, where excess activity of the CXCR4/CXCL12 system retain neutrophils in the BM niche, leading to enhanced neutrophil apoptosis (ie, myelokathexis) and NP.^42^ A similar mechanism may hypothetically occur in AIN, related to local mechanisms, for example, autoantibodies to epitopes on maturing/matured neutrophils.^36^ However, increased BM neutrophil apoptosis has not been clearly described in AIN. Considering that an estimated number of 10^9^ neutrophils/kg of body weight are formed daily in a healthy adult, at least a similar number would have to undergo apoptosis in the BM of AIN patients.^43^ However, BM biopsies most often appear normal in AIN or display myeloid hyperplasia (perceived as a sign of increased neutrophil production due to supposed increased peripheral destruction).^18^ This hypothesis implies the existence of a feedback system, resulting in a compensatory increased neutrophil production; however, no evidence for such a chain of events has been demonstrated so far in AIN.

Findings of elevated hCap-18/pro-LL-37 blood levels might support a hypothesis of increased neutrophil production in AIN resulting from a positive feedback loop from peripheral destruction. Conversely, reduced blood levels may indicate destruction of mature neutrophils in the BM or the peripheral tissues.^44^

The second removal site is considered to be the spleen. It has been assumed that AIN is mirroring what happens in autoimmune hemolytic anemia, where sequestration of autoantibody-coated red cells in spleen is documented. Imaging techniques with tagged neutrophils have so far not been conclusive because the labeling procedure may damage the cells, precipitating premature cell death. Thus, such findings may not answer the question. There is limited information on the role of splenectomy (eg, for Felty syndrome) on resolution of AIN; although ANC may rise following splenectomy, this may be due to removing a large bulk of active immunological tissue.^13^

The next possibility to explain neutrophil clearance is that large numbers of these cells, perhaps tagged by autoantibodies, return to the BM for destruction. This pathway for elimination of healthy aged neutrophils has attracted considerable interest.^39–41,43,44^ However, in a recent study, increased apoptosis of PB neutrophils did not affect the blood ANC.^45,46^ Anecdotal demonstration of phagocytosis/emperipolesis of neutrophils bymacrophages in the AIN BM needs quantitative data to explain a massive cell removal in AIN.^47^

How are neutrophils removed in AIN?

Thus, although autoantibodies to neutrophils have been considered as the main mechanism of neutrophil destruction in AIN, this has so far not been convincingly demonstrated. Adding to this conundrum, the emerging understanding that some AINs are not associated with detectable autoantibodies (but other phenomena of autoimmunity) has prompted a discussion of alternative mechanisms.

The concept of immune recognition and response to cell damage and death is rapidly increasing. One example of potential interest for AIN is myeloperoxidase-antineutrophil cytoplasmic antibody (MPO-ANCA)-associated vasculitis.^48,49^ Briefly, in this disorder cell surface bound/exposed MPO may activate the immune system, that subsequently activates the neutrophil, leading to a special form of degranulation and generation of neutrophil extracellular traps (NETs). During NETosis, the exposure of previously protected granule constituents to neighboring immune cells result in autoimmune reactions with development of antibodies to these peptides.^49^ It is possible that the autoantibodies serve as opsonins for enhanced phagocytosis or emperipolesis of neutrophils by macrophages in the spleen or BM.

Therefore, the question is if the HNA-directed or other neutrophil-directed autoantibodies cause the NP. It is reasonable to assume that this happens, but demonstration of a cause-effect relationship is needed.

Redistribution as a cause of blood NP

Acute redistributions of neutrophils leading to NP are well known, for example, after endotoxin exposure and acute complement activation; the lungs may be the site of neutrophil accumulation in these cases.^40,41,43,45^ Less is known of chronic redistribution as a cause of NP. The recent description of the possibility that NP in ADAN is due to a shift of the mass of neutrophils from the PB to the tissues, including the spleen, is a reminder of the need for better estimates of the total body neutrophil count.^50,51^

Neutrophil subpopulations

Finally, some unresolved, yet interesting aspects of the pathogenesis of AIN remain. For example, it is unknown if certain subpopulations of neutrophils are specific targets of immune mechanisms, for example, those expressing CD177 or those interacting with structures on endothelial cells significant for transmigration to the tissues or from the tissues back to PB or BM, and those interacting with organ-specific epitopes.^52^ Likewise, it is unknown if there are recognition signals of neutrophil subpopulations for interaction with cytotoxic T-lymphocytes, NK cells, monocyte/macrophages, MDSC, and other cells that are involved in elimination or cross-talks with neutrophils.^15,17,39,45,53,54^

Other possible mechanisms for AINP

Cellular immune mechanism, implicating clonal or oligoclonal expansions of T-cell and NK-cell populations, may also account for NP particularly in adults.

The 2 representative models of cell-mediated NP are the “LGL lymphoproliferative neoplasms” and the chronic unexplained, usually benign NP, also known as CIN.

The clonal expansions of LGL population

They are also known as LGL leukemias and are currently classified by the World Health Organization as T-LGL leukemia, chronic lymphoproliferative disorder of NK cells, and aggressive NK-cell leukemia.^55^

The first 2 entities are the most common among these neoplasms and display a similar, rather indolent course; the latter is rare, it is mostly found in Asia and is association with Epstein-Barr virus infection. STAT3 mutations and epigenetic changes are common in LGL leukemias of both T- and NK-cell types, whereas STAT5b mutations are uncommon and usually associated with an aggressive disease.^13,56–59^

The pathogenesis of NP in LGL leukemias is multifactorial, implicating humoral and cytotoxic mechanisms. In addition, neutrophil sequestration in the spleen and/or defective neutrophil production due to BM infiltration by the abnormal LGLs represent contributory mechanisms.^13,14,60–62^

Antineutrophil autoantibodies and circulating immune complexes have been identified in patients’ serum and their possible implication in immune-mediated neutrophil destruction has been postulated.^60^ However, the BM examination in LGL leukemia usually reveals mild hypercellularity with left-shifted neutrophil forms and decreased neutrophil precursors, suggesting defective proliferative/survival characteristics of cells of the neutrophil lineage.^14,60^ Indeed, the expanded, STAT mutated LGLs display overactivation of Ras–Raf-1–MEK1-ERK, PI3K/Akt, NF-kB, and sphingolipids pathways.^14^ Thus, they exhibit a survival advantage, overexpress, and secrete FasL and produce proinflammatory cytokines such as IFN-γ, IL-8, IL-10, IL-1β, IL-12p35, IL-18, IL-1RA, RANTES, MIP1-α, MIP1-β that inhibit the proliferation and/or induce apoptosis of the BM granulocytic progenitor/precursor cells and PB neutrophils. They also release perforin and granzymes that may further suppress neutropoiesis through a direct cytotoxic mechanism.^60^

Chronic idiopathic neutropenias

Activated, oligoclonal cytotoxic T-cell expansions, identified by flow cytometry, T-cell receptor beta chain (TRB) complementarity-determining region 3 (CDR3) PCR spectatyping and sequencing analysis, are detected in a significant proportion (60%–80%) of patients with CIN. These cell expansions do not fulfill the criteria for T-LGL leukemia; however, they display myelosuppressive properties by producing proinflammatory cytokines and proapoptotic mediators such as TNF-α, IFN-γ, and FasL that result in the apoptotic death of the BM CD34^+^/CD33^+^ neutrophil progenitor cells.^14,63^ The sequencing analysis of the TRB CDR3 has also revealed shared clonotypes between different CIN patients, possibly suggesting a common underlying antigen stimulation.^64,65^

However, not only T-cell expansions but also myeloid cell populations may contribute to the immune processes associated with CIN. It has thus shown that the monocyte cell subsets are altered in CIN patients and consist of decreased number of classical CD14^bright^/CD16^−^ and increased intermediate CD14^bright^/CD16^+^ and nonclassical CD14^dim^/CD16^+^ cells.^53^ The latter cell populations display increased expression of proinflammatory cytokines and higher potential for antigen presentation.^53,65^ Finally, another interesting cell population with immune-regulatory properties are the MDSCs.^15^ MDSCs are identified as CD11b^+^CD33^+^HLA-DR^−/low^ cells and further characterized as CD14^+^ (monocytic-MDSCs, M-MDSCs) or CD15^+^ (polymorphonuclear-MDSCs, PMN-MDSCs) cells.^66,67^ As regards to their functional properties, they are mainly characterized by the capacity to suppress T-cell activation and proliferation through multiple mechanism. MDSCs have emerged as important contributors to immune dysregulation associated with autoimmune and chronic inflammatory disorders, (hematologic) malignancies and immune-mediated cytopenias such as thrombocytopenia.^15^ Their role in AIN is currently under investigation.

TYPES OF NPS

pAIN and secondary AINs

In most classifications on AIN, this disorder is designated as primary or as secondary. pAIN is a common NP type which arises in early infancy, usually before the age of 3 years and shows a mild/moderate clinical course. It resolves spontaneously within 24–36 months of onset in 90% of cases.^8–12,68,69^

It has been assumed that pAIN is rare in adults but data on this entity are scarce (Figure 2). Conversely, secondary AIN (sAIN) is mostly seen in adults, is prevalent in females, has low tendency to remit^17,70,71^ and is usually associated with other autoimmune phenomena or diseases (ie, Felty syndrome, systemic lupus erythematosus [SLE], scleroderma, Sjögren syndrome), malignancies (eg, LGL leukemias, various lymphomas), acute and chronic viral infections (eg, HIV, hepatitis, cytomegalovirus), drug reactions, and after hematopoietic stem cell transplantation (HSCT).^17,72,73^

AIN in children and adolescents

The most common type of NP in pediatric age is the pAIN which is classically diagnosed early in life and remits in the majority of the affected subjects.^9^

Primary idiopathic NP is another common NP of the first years of life; the disease phenotype overlaps with the one of pAIN, from which it differs because of no detection of antineutrophil autoantibodies even on repeated tests over time. Considering the nonoptimal sensitivity of antineutrophil antibody assays, as described above, there seems to be general consensus that primary idiopathic and pAIN are, in the majority of children, the same disease.^33^

In a minority of pediatric patients, AIN has different characteristics from the classical pAIN. The features that make this kind of NP different (i) long more than the typical 24–36 months and (ii) late appearance, after the usual age of 3 years. Patients affected with late onset or long-lasting NPs may sometimes be diagnosed by chance, due to the very mild clinical phenotype and do not usually show remission. They also differ from sAIN since they are not generally associated with underlying autoimmune signs/disorders.^70^

Recently, a study from the Italian Registry of NP identified this category of long-lasting NP and NP of late onset as different from classical pAINs and sAINs.^74^ This new category, indeed, displayed also some distinct features such as lymphopenia with decreased B and NK cells numbers. Notably, genetic analysis showed that some of these “atypical” NP patients carried pathogenic variants that were considered causative of immunodeficiency and immune dysregulation disorders of which NP was the most evident feature.^72^ These findings have clinical relevance since they suggest that pediatric diagnostic work-up and management of NPs lasting longer than the typical time frame and appearing in late childhood and adolescence should be fairly different than the typical pAIN diagnosed in early infancy or remitting within the expected time frame. One might speculate that NPs, arising in late infancy or adolescents, are paucisymptomatic and possibly underdiagnosed. They might not be detected until adulthood and present then as CIN or AIN (Figure 3).

AIN in relationship to other autoimmune phenomena and disorders

pAINs are occasionally seen in adults; this diagnosis is based on the presence of autoantibodies and the absence of any underlying disorder that may cause NP. The disease is rarely complicated by infections, the NP is mild to modest and tend to persist for years, even decades. The relation to immunodeficiency syndromes, as for the childhood disorder described above, is unclear.

More common are the sAINs.^11,12,17^ Patients with sAIN, mostly females 25–40 years of age, may present with various symptoms that resemble classic autoimmune disorders (Figure 3). Apart from arthralgias (but rarely arthritis), solar intolerance, Raynaud phenomena, excessive fatigue, and sometimes skin rashes, these NP patients may also present with positive screening tests for various antinuclear antibodies, serum polyclonal immunoglobulin G or M (less often IgA) rises, presence of autoantibodies to thyroid, liver or biliary tissues, celiac disease, and/or signs of serum complement activation. Many but not all present with autoantibodies to HNA antigens. However, despite an extensive work-up, many do not reach the established criteria for SLE or other classic autoimmune diseases.^75^ Their symptoms are often fluctuating, with remissions and recurrences, that occasionally coincide with ANC variations. It is unclear if they share the genetic signature displayed by those with a definite rheumatic/autoimmune diagnosis.^76^

The BM examination yields most often unremarkable results. The plasma level of pro-LL-37 is often normal; it remains to show if it is higher than normal, indicating an enhanced myelopoiesis.^18^

The distinction between AIN and CIN in the adult NP patient is not always clear and depends largely on the persistence of the treating physicians in repetitively performing the antineutrophil antibody testing. Thus, the NP diagnosis may be changed between CIN and AIN with time. Interestingly, in a cohort of severe AIN/CIN patients studied as a single group on the basis of the diagnoses assigned by the referring physicians, 3% developed a lymphoid malignancy, whereas no propensity toward a myeloid malignancy was identified.^77^ In contrast, a recent study evaluating CIN patients with no evidence of antineutrophil antibody activity, an increased risk for the development of myeloid malignancies was identified in patients with clonal hematopoiesis.^78^

The remission rate for AIN in adults is unknown. In the authors’ experience, rare cases of remission occur, but long follow-up studies are required for estimation of frequency as well as the risk for progression to hematologic malignancies.^17,77^

The remission rate for AIN in adults is unknown. In the authors’ experience, rare cases of remission occur, but long follow-up studies are required for estimation of frequency as well as the risk for progression to B-lymphocyte malignancies.^77^

AIN as part of primary immune deficiency syndromes

The most common primary immune deficiencies (PIDs) associated with NP are X-linked agammaglobulinemia, hyper-IgM syndrome and WHIM syndrome, which are usually diagnosed early in life (Table 2). In these disorders myelopoiesis is either suppressed due to defective intrinsic mechanisms of maturation or shows a poor/absent egress of mature cells from the BM, as seen in the WHIM syndrome.^79,80^

Less commonly, some PIDs appear late in life, in adulthood; this is the case of common variable immunodeficiency (CVID).^81^ It has been postulated that in patients affected with CVID, the progressive loss of Treg cell function together with a severely reduced number of switched memory B cells allow the emergence of autoreactive clones which are the cause of autoimmune cytopenias.^80–82^

An important “alert” for the hematologist is that a non-negligible portion of patients diagnosed with autoimmune cytopenias, including AIN, are affected with an underlying immunodeficiency which may fully develop later on (Table 2).^82^

Special forms of AINs

NP in Evans syndrome

AIN may occur as a part of Evans syndrome, although less frequently than the combination of autoimmune hemolytic anemia and immune thrombocytopenia. NP may represent an early manifestation of this syndrome.^83,84^

Even though PIDs are most often diagnosed in children, they should also be considered in adults in cases with suggestive clinical presentation (ie, consanguinity, family history of autoimmune cytopenias or PID, recurrent infections, lymphoproliferation, lymphomas, hypogammaglobulinemia, or polyclonal hypergammaglobulinemia).^84–88^ Improved knowledge and diagnostic methods of genetic disorders facilitate the recognition of these cases.^83,84^

Autoimmune neutropenias and HSCT

Autoimmune disorders, and in particular autoimmune cytopenias, may follow allogeneic HSCT, with an estimated incidence of ~3% in adults and ~5% in children.^88,89^

The risk factors for the development of autoimmune disorders after HSCT a nonmalignant primary diagnosis; the use of an unrelated donor; the absence of total body irradiation in the preparatory regimen; the use of T-cell depletion by means of antithymocyte globulin and alemtuzumab in the conditioning regimen; the presence of chronic graft-versus-host disease (GvHD); and the use of peripheral or cord blood rather than BM as stem cell source.^88–91^

The proposed mechanism for emergence of post-HSCT autoimmunity might be breakdown of peripheral immune tolerance due to lack of functional T-cells, preferentially eliminated by conditioning regimens. In particular, the down-regulation of the Tregs favors B-cell expansions, leading to the autoimmune cytopenias. In addition, chronic GvHD may act as a risk factor due to inability of Tregs to dampen alloimmunity.^88–91^

The management of AIN and other cytopenias after HSCT is rather challenging. It includes the use of corticosteroids, rhG-CSF, and anti-CD20 monoclonal antibodies in refractory cases.^88–91^

Focus on follow-up

Longitudinal follow-up in young children is needed to verify the remission of pAIN. According to the available data in the Italian Registry, pAIN remissions occur in almost 95% of affected patients.^9^ Sometimes, resolution may occur abrupt. These “sudden” remissions were documented in almost 70% of the patients, whereas transient/intermittent oscillation (just above and below the lower reference value) was observed in the remaining 30% of cases.^9^

As for longitudinal follow-up of NP in late infancy and young adulthood, particularly when NP lasts more than the predicted time, it is highly recommended to monitor the patients for the appearance of systemic autoimmune symptoms and disease. This is true also for adults whose NP may have flares and remissions overtime.^77^ AIN adolescents and adults may have a chronic course that rarely progress toward severe NP. In some, it may disappear after a couple of years or present alternating flares and remissions. The annual monitoring of these patients, sometimes in consultation by a rheumatologist, is suggested to determine if antirheumatic treatment may confer amelioration of symptoms.^75^

In general, the risk to acquire bacterial infections is low in most AIN patients with mild/moderate NP. However, in those with severe NP, infection propensity may necessitate prophylactic use of antibiotics and/or rhG-CSF. Tools to assess infection propensity beyond the ANC are warranted.

TREATMENT OF AINS

The treatment of AIN is not much different from treating other NPs. The indication for treatment is the observed infection propensity (ie, infection frequency, severity, need for antibiotics, hospital admissions, patients’ choices, etc). It should be observed that the presence of a raised inflammatory indicators (ie, CRP), gingivitis, periodontitis, oral ulcerations, persisting despite good oral hygiene, are important to include in the evaluation. The ANC, per se, is not an indication for treatment, since many AIN (as well as idiopathic NP) patients may display severe NP without increased susceptibility to bacteria.

The most important pharmacological treatment for AIN both in adults and in children is prophylactic or therapeutic rhG-CSF (both fil- and lenograstim).^32,77,92^ The G-CSF treatment may be given prophylactically on long-term basis, or on demand. The later might be used to enhance resolution of severe or complicated infections. If an AIN patient faces planned surgery, particular one where postoperative infections might be devastating, it is advisable to temporarily increase the ANC, usually by means of rhG-CSF administration a few days before and after surgery.

Doses from 1 to 3 μg/kg//d in children or otherwise 12 MIE/0.2 mL in adults (or higher), 1–7 days a week, depending on the ANC, side effects and patient’s wishes, are advised.^18,20,32^

The ANC sample should be obtained the same morning as the subcutaneous injection is usually given in the evening, to assess the lowest ANC value, as the tool for decisions. Often, a nadir value of 1–1.5 G/L is sufficient for infection control, but the dose can be adjusted if break-through infections occur to reach higher ANCs. The treatment can be paused occasionally to see if NP still exists. The effect of the treatment should include assessments of oral health.

Common side effects of G-CSF treatment include bone pains, malaise, headache, often encountered during the ensuing day. If side effects still are a problem, we recommend lowering the G-CSF dose and accepting lower ANC but carefully monitor infection propensity. If G-CSF is administered 2–4 times a week, side effects are often reduced if daily injections with smaller G-CSF doses are used instead, but with maintained weekly cumulative dose.

A very unusual side-effect of G-CSF is a situation with fever, chills, and marked malaise 1–3 days after an injection. This reaction is most often seen at the very start of treatment. Avoiding prefilled syringes and changing to G-CSF preparations in vials, from which the patient withdraws the daily dose, may prevent those side effects.

Rarely, G-CSF might elicit an exacerbation of an underlying autoimmune disorder, for example, in the kidney, skin (Sweet syndrome), and other organs.^93^

The evolution to MDS/acute myeloid leukemia, described in a variable proportion of severe congenital NP patients, has not been described in AIN.^94–97^

Challenges in treatment

Patients carrying autoantibodies to CD177 have been reported to be resistant to treatment with filgrastim, but the biochemical basis for this observation is still unknown.^52^

Various immune-regulating drugs (eg, cyclosporine, metothrexate, low-dose cyclophosphamide) have been used in refractory cases, but overall efficacy has not been determined. In case of an underlying immune dysregulation/deficiency, particularly in childhood, the definition of the causative variants is critical because it may indicate more targeted treatments (eg, mycophenolate mofetil and rapamycin).^98,99^

The role of rituximab (and other antibodies to CD20) is well established as effective treatment for many autoimmune disorders. However, its role as treatment for AIN is unclear.^100^

Corticosteroids, despite being efficient in the short-term perspective, and intravenous/subcutaneous gammaglobulins are rarely needed, but for emergency situations.

Prospects for better diagnostics and personalized treatment

As discussed in this review, AIN may comprise a variety of distinct subclassifications, that differ with regard to presence/absence of autoantibodies to neutrophil HNA epitopes, genetics, clinical course, and prognosis. We have discussed the importance of cellular mechanisms in addition to humoral effects that impact the survival of the neutrophils and/or their progenitors resulting in the AIN phenotype. The delineation of the humoral/cellular mechanisms and the identification of the genetic background of the atypical forms remains a challenge for a more precise and individualized treatment approaches.

AUTHOR CONTRIBUTIONS

All authors contributed to the writing of this review.

DISCLOSURES

FF and HAP performed an advisory board for X4 Pharmaceutical; JP did a Consultancy for Chiesi Canada Ltd; CD has disclosures with Pfizer, Gilead, Novartis, Rockets, and Bokryst; the other author has no conflicts of interest to disclose.

SOURCES OF FUNDING

The study is based on work from COST Action CA18233 “European Network for Innovative Diagnosis and treatment of Chronic Neutropenias, EuNet-INNOCHRON” supported by COST (European Cooperation in Science and Technology).