Homeostatic regulation of blood neutrophil counts

Abstract

Neutrophil counts in blood are determined by the differentiation and proliferation of precursor cells in the bone marrow, release of mature neutrophils into the blood, margination in organs like the lung and spleen, and transmigration through the endothelial lining followed by neutrophil apoptosis and uptake by phagocytes. This brief review summarizes how the regulation of neutrophil production by G-CSF is in part controlled by IL-17 and IL-23. Neutrophils are retained in the bone marrow through interaction of CXCL12 with its receptor CXCR4. The relevance of this mechanism is illustrated by rare diseases in which disrupting the desensitization of CXCR4 results in neutrophil accumulation in the bone marrow. Although blood neutrophil numbers in inbred mouse strains and individual human subjects are tightly controlled, the large variation of blood neutrophil counts among outbred populations suggests genetic control. One example is benign ethnic neutropenia, which is found in about 5% of African Americans. Reduced and elevated neutrophil counts, even within the normal range, are associated with excess all-cause mortality.

Introduction

Neutrophilic granulocytes (neutrophils), the most abundant, but also very short-lived human white blood cells, act as fast primary defense against infections (1). Neutrophil turnover is rapid - around 10⁹ cells/kg body weight leave the bone marrow per day in healthy humans (2, 3). Bone marrow postmitotic transit time determined by maximal blood neutrophil radioactivity after a pulse of ³H-thymidine was 7 days (2, 3). The transit time in rabbits and mice was somewhat shorter: the peak of cell mobilization into peripheral blood occured about 95 hours after leaving the mitotic pool, where progenitors remained about 50 hours (4, 5). Within the circulation, the half-life of infused, radiolabeled neutrophils was 7–10 hours in humans (3, 6) and 11.4 hours in mice (4). In rabbits, a shorter half-life of 3.2 hours was reported (7, 8).

Neutrophil progenitor proliferation and differentiation

Hematopoietic cytokines promote neutrophil progenitor proliferation and differentiation acting in a complex network (9). The major cytokine for neutrophil proliferation and survival is G-CSF. Mice and humans deficient in either G-CSF or its receptor suffer from profound neutropenia (10–12). G-CSF currently is the major therapeutic agent for neutropenia of iatrogenic as well as genetic and various other origins (13–15). Extensive basic science and clinical data exist on the role of other granulopoietic cytokines such as M-CSF, GM-CSF, interleukin (IL)-6, IL-3, IL-17 and, most recently, IL-22 (10, 16–22) that have been reviewed elsewhere in detail (23). Genetic modification of intracellular messengers downstream of G-CSF in mice elucidated their stage-specific roles (24). For example, both STAT3 and SOCS3 deficiency resulted in neutrophilia and an increased pool of late stage progenitors in the bone marrow thus implicating an inhibitory role (25–29). The role of transcription factors and microRNA in neutrophilic differentiation has recently been reviewed (30, 31). A number of monogenic defects associated with rare forms of congenital neutropenia in humans are known. Maturation arrest and increased cell death of neutrophil progenitor proliferation have been observed in humans with elastase gene mutations, but also in genes encoding a number of transcription factors such as Growth factor independent 1 (GFI 1), HCLS1 associatied protein X-1 (HAX1), and lymphoid enhancer factor-1 (LEF-1) (32).

Neutrophil mobilization from the bone marrow

Release mechanisms of hematopoietic stem cells, myeloid progenitors and granulocytes from the bone marrow have been studied extensively under normal and emergency conditions (33–35). G-CSF is a most active soluble physiological factor. In regard to neutrophils, the interaction of stromal derived factor-1 (SDF1, CXCL12) with the chemokine receptor CXC-receptor 4 (CXCR4) is important for retention in the bone marrow. CXCR4 deficiency resulted in decreased bone marrow but increased peripheral neutrophils as identified by the marker Gr-1 (36). CXCR4 and CXCL12 are down-regulated by G-CSF (37, 38). Neutrophil mobilization can be induced by anti-CXR4 antibodies and a number of peptide antagonists (39, 40). Conversely, activating mutations of CXCR4 in humans cause neutrophil accumulation in the bone marrow together with peripheral neutropenia together with a complex immunological phenotype (WHIM syndrome; warts, hypogammaglobulinemia, immunodeficiency, myelokathexis) (41–43). This year, specific patients’ mutations have given further insights into downstream signaling: in one WHIM patient decreased expression of the G-protein coupled receptor kinase (GRK) 3 was observed. GRKs are essential for desensitization of CXCR4 and subsequent neutrophil release from the bone marrow (44). Another mutation inhibited internalization of phosphorylated CXCR4 (45).

Neutrophil serine protease expression correlates with neutrophil release from the bone marrow. Cathepsin G and neutrophil elastase, but also matrix metalloproteinase 9 were increased by G-CSF treatment. Inhibition by alpha-1-antitrypsin inhibited neutrophil release from the bone marrow (46–48). However, neither deficiency in both cathepsin G and neutrophil elastase nor a mouse model lacking the serine proteinase activator dipeptidylpeptidase I showed altered neutrophil mobilization, thus challenging the role of serine proteases in neutrophil liberation (38).

Margination, adhesion and migration into tissues

In mice, the circulating pool of neutrophils amounts to only 1–2% of the morphologically mature neutrophils in the bone marrow (49). Neutrophil homing studies have mainly depended on extracorporally labeled cells. Data have to be interpreted cautiously as partial cell activation as occurs during isolation alters homing properties (50). In one study, approximately a third of reinfused neutrophils was found in liver and bone marrow and about 15% in the spleen, but the target depended on the collection method. Neutrophils from thioglycollate-induced peritonitis preferentially homed to the liver and bone-marrow neutrophils to the bone marrow when assessed after 4 hours (51). Endotoxin- or cobra venom factor-mobilized neutrophils infused into rats were found at 21% in spleen, 22% in liver and 14% in lungs after 4.5 hours (52). The vasculature of the lung harbours a considerable neutrophil pool. In rabbits, about 20% of ⁵¹Cr labeled neutrophils stayed in the healthy lung, of those around 90% in capillaries (53). Catecholamines can mobilize marignated neutrophils. Interestingly, altered mobilization of marginated neutrophils may be a factor in ethnic neutropenia in humans: in addition to low baseline counts, affected subjects mobilized fewer neutrophils during marathon running or other strenuous exercise (54).

Integrins and selectins are essential initiating neutrophil exit from the blood pool (55, 56). Specific adhesion molecule deficiencies increase circulating neutrophil numbers. Mice deficient in leukocyte function associated antigen (CD11a, Itgal^−/−) or the common chain of all β₂-integrins (CD18, Itgal^−/−) show marked leukocytosis (57, 58). Neutrophil migration to various tissues was reduced in Itga^−/−-deficient mice (58, 59). Itgal-silencing by neutrophil specific microRNA recently confirmed this phenotype (60). Mild neutrophilia was also found in mice deficient for P-selectin (Selp^−/−) (61, 62) and more marked if both E- and P-selectin (Selp^−/− Sele^−/−) or all selectins (Selp^−/− Sele^−/− Sell^−/−), were absent (62, 63). Absence of and enzyme required for selectin glycosylation, core 2 beta-1,6-N-acetylglucosaminyltransferase (Core2^−/−), resulted in neutrophilia (64).

Neutrophilia in adhesion molecule-deficient mouse strains was initially thought to be caused by passive neutrophil accumulation in blood vessels. To test whether adhesion-molecule-deficient neutrophils accumulated more than wild-type cells, several groups used mixed ^Itgal−/− and wild-type bone marrow transplants into wild-type mice. Surprisingly, the percentage of wild-type and Itgal^−/− neutrophils in peripheral blood and in bone marrow was very similar (65–67). Even a small proportion of wild-type cells was sufficient to normalize blood neutrophil levels. Proliferation measured by BrdU incorporation of Gr1 positive bone-marrow cells did not differ between wildtype and Itgal^−/− cells six months after transplantation. This argues against intravascular accumulation or autonomous proliferation as reasons for neutrophilia in Itgal^−/− mice.

In humans, leukocyte adhesion deficiencies (LAD), rare diseases caused by either deficiency or signaling dysfunction of β₂-integrins (LAD I), selectin ligands (LAD II), or downstream signaling molecules (LAD III) replicate the neutrophilic phenotype of the respective gene-deficient mice (68–70).

Apoptosis, clearance and feedback

Neutrophils rapidly die by apoptosis in the absence of external stimuli. When apoptosis was induced in vivo, neutrophils were mainly found in the liver, where they were phagocytosed by Kupffer cells (71). Apoptotic neutrophil phagocytosis has an anti-inflammatory role (72). The proinflammatory cytokine IL-23 consists of a p40 and a specific p19 subunit. It is induced in macrophages by a variety of transcription factors, including NFkappaB, that can be downregulated by neutrophil phagocytosis (73, 74). Transgenic overexpression of the IL-23 specific subunit p19 in mice induced neutrophilia (75). Conversely, IL-23 deficiency or blockade with an antibody decreased neutrophil counts in normal and neutrophilic mice (76).

IL-23 is a potent inducer of IL-17, the most prominent member of a cytokine family defining the T_H17-CD4-subpopulation (77). In all strains of severely neutrophilic, adhesion molecule-deficient mice, elevated IL-17 levels were found, and IL-17 blockade by soluble IL-17 receptor demonstrated that their neutrophilia was indeed caused by IL-17 (65). Mice deficient in the IL-17 receptor (Il17ra−/−) had decreased neutrophil counts (19, 78). IL-17 stimulates G-CSF secretion (79), and G-CSF levels were elevated in all neutrophilic mouse strains, where blockade of G-CSF normalized neutrophil counts (65). Closing this feedback loop, IL-23 expression in peripheral tissues was reduced by apoptotic neutrophils (59). These data suggest a model where granulopoiesis is driven by a cytokine cascade starting with macrophage and dendritic cell IL-23 secretion. The resulting in T-cell IL-17 secretion increases G-CSF levels. When neutrophils arrive in peripheral tissues, their phagocytosis downregulates macrophage IL23 secretion and - via decreased IL17 and G-CSF- curbs granulopoiesis (80).

“Normal” neutrophil count

Baseline neutrophil counts are relatively stable in individuals but have a considerable normal range in healthy humans. A survey of more than 25,000 Americans found a mean neutrophil count of 4.3×10⁹/l in adult males and 4.5×10⁹/l in females for Caucasian participants (81). In addition to environmental factors, whose influence was highlighted by a recent study showing a global decrease of neutrophil counts in an US-Amercian population from 1958 to 2002 (82), the genetic background is important. Mean neutrophil counts are lower in African Americans, in one study 3.5×10⁹/l in males and 3.8×10⁹/l in females (Fig. 1a)(81). “Benign ethnic neutropenia” is a condition found in up to 5% of African Americans and is defined as a neutrophil count below 1.5× 10⁹/l without overt cause or complication (81, 83). Little is known about the genetic factors that influence this difference or human steady state granulopoiesis within the normal range. Variation was also seen between different inbred mouse strains. Neutrophil counts from four commonly used mouse strains are given in figure 1b. (84, 85). Whole genome association studies in F2 intercrosses in mice and swine revealed chromosomal regions associated with blood neutrophil counts (86–88), some of them harbouring coding regions for cytokines such as IL-2, IL-15, IL-12 and chemokines such as CXCL8. However, specific mutations leading to functional alterations of these cytokines remain to be determined.

Figure 1. Normal range of neutrophil counts in humans and mice.

A) Neutrophil counts from 25,000 US Americans from (81) modified to show cumulative incidence. Mean counts in African Americans were significantly lower than in Caucasian or Hispanic individuals.

B) Neutrophil counts in inbred mouse strains. Neutrophil counts calculated from white blood counts and relative neutrophils counts from 129S1/SvlMJ (n=29), BALB/cJ (n=16), FVB/NJ (n=24) and C57BL6/6J (n=19) from the Jackson laboratory phenome database (84, 85).

Clinical relevance of baseline neutrophil counts

Neutrophilia is a classical indicator of acute inflammation, while idiopathic and acquired forms of neutropenia predispose to infections (89). However, total white blood counts (WBC), that are mainly determined by neutrophil counts in healthy humans, are also relevant in the absence of acute events. Increased WBCs have long been associated with increased all-cause mortality (90–94). A prospective study over 44 years revealed a j-shaped association curve of neutrophil, but not lymphocyte count and all cause mortality (82) (Fig. 2)

Figure 2. Relationship between excess mortality and white blood count.

Nearly 4000 individuals from the Baltimore/Washington area were observed from to 1958–2002. Excess mortality as difference between observed and expected mortality hazard over time is plotted against white blood cell count. The dashed lines represent the 95% confidence intervals (with permission from (82).

Neutrophils are the first defense against invading microorganisms. Increased susceptibility to common pathogens has usually been attributed to extremely low counts (below 0.5×10⁹/l) (13) and individuals with “low normal” counts or ethnic neutropenia have not been reported to be at increased risk as long as counts are not further decreased. However, the probability to contract tuberculosis from patients with open pulmonary disease was inversely correlated with baseline neutrophil counts in a recent study (90). In contrast, an increased total WBC and neutrophil count has been shown to be an independent risk factor for cardiovascular mortality in a number of studies and subsequent metanalyses (82, 92, 95–99). Less data exist on neutrophil counts and cancer mortality, but data from the NHANES population show a higher total WBC as an independent risk factor for total cancer mortality (100). However, it is at present unclear if the elevated numbers of circulating neutrophils are causative of the observed increase in mortality or rather a measure of ongoing subclinical inflammation (101).

Summary

Stable neutrophil blood counts are the result of a highly dynamic feedback system. The study of genetically altered mice and monogenic diseases in humans has given insight into some of the involved mechanisms. However, neutrophil counts in healthy humans are regulated by a variety of environmental and genetic factors, most of which remain currently unknown. As elevated counts within the normal range are associated with excess mortality, elucidation of factors involved in steady state neutrophil regulation might have clinical relevance.