The proliferative human monocyte subpopulation contains osteoclast precursors

Rheumatoid arthritis (RA) is a chronic disease that is characterized by joint inflammation and profound focal and generalized bone loss due to the action of osteoclasts [1, 2]. Multinucleated osteoclasts derive from monocyte/macrophage lineage precursors; two key mediators controlling their development are macrophage colony-stimulating factor (M-CSF or CSF-1) and receptor activator of nuclear factor-kappa-B ligand (RANKL) [3‐5]. Human osteoclast precursors have been shown to be present at low frequency in normal peripheral blood [6‐8]. It now appears that peripheral blood monocytes, which derive from bone marrow precursors, are heterogeneous as judged by criteria such as surface marker expression, size, and function [9]. For example, in the human, there is a minor subpopulation of monocytes which is CD14^lo CD16⁺ [10] and which has been implicated in inflammation and cancer [11‐13]; in the mouse, there is a lot of recent interest in monocyte subpopulations that appear to have different roles during inflammatory reactions as manifested, for example, by their ability to migrate to sites of inflammation [14]. Human osteoclast precursors have recently been shown to reside in the CD14⁺ CD16^- monocyte subpopulation of normal donors [15]. Blood samples from psoriatic arthritis patients, particularly those with bone erosions visible on plain radiographs, exhibit an increase in osteoclast precursors compared with those from healthy controls [16]; these precursors were recently reported to reside in the CD14^lo CD16⁺ monocyte subset, leading the authors to suggest that osteoclasts are derived from distinct monocyte subsets in these patients and in healthy individuals [17].

Human monocytes are commonly considered to be non-proliferating [18]; however, we and others have defined a subpopulation of human monocytes which is capable of proliferating in vitro (for example, in response to M-CSF) [19‐25]. This population has been referred to as proliferative monocytes (PMs), which were shown recently to have the phenotype CD14⁺ CD16^- CD64⁺ CD33⁺ CD13^lo c-Fms⁺ prior to culture [25]. It was previously suggested that PMs might be able to migrate into inflamed tissues and possibly undergo local proliferation there [19, 25]. During these prior phenotyping studies, it was noticed in passing that, following culture and sorting by flow cytometry, the PMs, from the few donors studied, could give rise to tartrate-resistant acid phosphatase-positive (TRAP⁺) multinucleated cells upon culture in M-CSF + RANKL [25]. Based on this preliminary observation and the likelihood that the PMs represent a relatively less mature monocyte population on account of their ability to proliferate, it was reasoned that they may retain differentiation capability and therefore contain the osteoclast precursors. We present evidence here for this concept for the peripheral blood from normal donors.

Under steady-state conditions, osteoclastogenesis and bone remodeling occur mainly in the bone marrow. Osteoclast precursors can be mobilized from bone marrow into blood and then into tissues, particularly in some conditions involving bone loss at diseased sites (for example, RA) [30, 31]. At an early stage of differentiation, they are also able to give rise to different myeloid populations [32], whereas at a later stage their differentiation involves M-CSF-dependent action on c-Fms⁺ populations [33].

The heterogeneity of peripheral blood monocytes has led to the concept that there may be distinct subpopulations of cells with specialized functions [10, 14, 34, 35]. For example, for the human, it is known that only a small proportion of monocytes can differentiate into osteoclasts [36, 37]. Likewise, it is known that a small proportion of CD14⁺ human monocytes (that is, PMs) can proliferate in vitro [19, 25]; because of their ability to proliferate, it was reasoned that this less mature population, possibly representing cells recently mobilized from bone marrow, may be able to differentiate into different macrophage lineage populations, such as osteoclasts, under appropriate conditions.

Taking advantage of the relative ability of monocyte populations to undergo proliferation, we were able to show above that, for the blood from 13 donors, osteoclast precursors reside predominantly in the PM population and could be detected even after proliferation in M-CSF. Following further culture in M-CSF and RANKL, the resultant population containing the multinucleated progeny showed increased expression of certain osteoclast markers (CTR, Cath K, and RANK) and an ability to resorb bone. These findings are consistent with the concept that the PMs represent a less mature population, when compared with the bulk of the human peripheral blood monocytes [19‐21], with some cells in the PM fraction at least retaining an ability to differentiate into osteoclasts. The data presented are consistent with prior observations that both the PM population [19, 25] and osteoclast precursors [15] from normal individuals reside in the CD14⁺ CD16^- monocytes rather than in the CD14^lo CD16⁺ population, that osteoclastic cells can be generated from proliferating dendritic cell precursors in human peripheral blood [38], and that there is an early increase in the percentage of human peripheral blood osteoclast precursors entering S phase during their in vitro differentiation in M-CSF + RANKL [39]. It is possible that the PMs can differentiate while in the blood into NP monocytes with reduced proliferative and differentiation potential, perhaps under the influence of circulating M-CSF.

It is intriguing that, in psoriatic arthritis, the opposite finding has been made in that the increased numbers of peripheral blood osteoclast precursors noted were located in the CD16⁺ population [17]. It would be worth knowing whether the PM population also increases in this and perhaps other inflammatory conditions and whether they begin to express higher CD16 levels in vivo. We suggest again [25] that functional criteria, such as PM status, have an advantage over surface marker phenotyping in that they avoid the difficulty in defining, for example, for monocyte populations whether modulation in the expression of a particular surface marker reflects differentiation or activation.

In summary, it has been shown here that human peripheral blood osteoclast precursors reside in the PM subpopulation, which is presumably a relatively less mature subpopulation and therefore possibly recently mobilized from bone marrow [19‐21]. It has been proposed before [19‐21, 25] that, upon migration into inflammatory lesions, the PMs may contribute to the local macrophage proliferation which can be observed [40, 41]. It is also possible that they could reside as osteoclast precursors in the synovial macrophage population within RA joints [30] which have been shown capable of differentiation into osteoclasts [42, 43].