In the present study, we demonstrate for the first time the in-vitro generation of FDC-like cells from monocytes as their putative precursors. Human monocytes were prepared by selective adherence combined with reverse sedimentation as a novel procedure in order to eliminate undesired non-adherent cell populations. This protocol allowed for a virtual enrichment of monocytes while avoiding any interference with antibodies posing the potential hazard to evoke adverse cellular trigger events.
Our data clearly demonstrate that peripheral blood monocytes can be induced to differentiatiate into FDC-like cells. Similar to the earlier conundrum on the ontogeny of T cell-tropic dendritic cells (DCs) [21
], the search for the origin of FDCs has always been complicated by their unstable marker profile, showing that FDCs isolated from lymphoid organs readily lose typical antigens such as DRC-1, CD21, CD23, and KiM4 in culture [35
]. For practical reasons, we therefore decided to take advantage of an AP isoenzyme that has been described to be expressed by FDCs [30
], but not by monocytes, macrophages or monocyte-derived DCs (cf. Tab. 1). Actually, our earlier experiments had revealed that this anchor marker is inducible in monocytes under defined conditions, suggesting AP expression as indicative for the developmental plasticity of these cells [27
]. As a logical step, when attempting to generate FDCs in vitro
, we first converted monocytes into APpos
cells, here referred to as primary FDC-like cells.
Typically, the common leukocyte antigen CD45 was strongly down-regulated in APpos
cells as well as in cells with similar morphology suggesting a transdifferentiation into non-leukocytic cells. Absence of CD45 has been described for FDCs by Schriever et al. [40
]. The co-expression of CD68, however, indicates the monocytic origin of the APpos
Labelling of CD35 (complement receptor I) is routinely used in histology to discriminate FDCs from other types of dendritic cells as well as from histiocytes [31
]. In our work CD35 was stably expressed in monocyte-derived FDC-like cells. Again, we found the monocyte/macrophage marker CD68 in CD35pos
cells which was also obvious in CD35/AP co-expressing cells. Control macrophages were stained only marginally or negatively for CD35.
When establishing conditions for the generation of APpos cells, sera were found to give variable results. Therefore, we developed a completely serum-free differentiation protocol. Low concentrations of IL-4 and GM-CSF plus Dex were found to act as basic inducers. Within 12 to 15 days of culture, these conditions elicited a cell showing markers and functions of FDCs. Conversely, this combination of factors prevented the differentiation of macrophages or T-cell tropic dendritic cells from monocytes.
IL-4 at concentrations one order of magnitude lower than those established for the differentiation of T-cell tropic DCs had previously been shown to generate large flattened and adherent cells with fine dendritic protrusions. Such cells had shown to express both, AP as well as the myeloid marker CD68 [27
]. Our forthcoming experiments revealed that IL-4 can successfully be replaced by IL-13, which will be a subject of further research.
Importantly, corticosteroids such as Dex synergistically enhanced AP expression. Dex, previously employed to induce osteoblasts from bone-marrow derived mesenchymal stem cells [41
], has been shown to inhibit macrophage and DC differentiation [42
]. When differentiating FDCs from monocytes, we now have found that IL-4 and Dex synergize in inducing the APpos
GM-CSF which was applied routinely at low concentrations appeared merely to serve as a survival factor in addition to its contribution to differentiation. Its actual requirement was largely donor-dependent, indicating that developmental properties of starter monocytes may vary according to the donor's immunological status. Moreover, we have experienced that occasionally exogenous GM-CSF can even be omitted without affecting the differentiation towards APpos
cells, which might be explained by an autocrine production of GM-CSF by monocytes in culture [43
After 12–15 days of culture, the majority of monocyte-derived cells displayed the APpos phenotype, while not proliferating throughout this period of time. This high percentage of APpos cells therefore reflects quantitative monocytic differentiation and clearly argues against the possibility that FDC-like cells might have expanded from a small contaminating cell population.
Alternatively, we used monocytes in which – as a first step – proliferation had been induced by adding GM-CSF at higher concentrations. Cells transferred from this condition and introduced to FDC-inducing conditions acquired the APpos phenotype as well. As a consequence, combining sequential proliferation and differentiation might emerge as the method of choice when considering large-scale production of FDC-like cells from monocytes, for example when faced with the need to avail such cells for therapeutic application.
Membrane expression of ICAM-I and VCAM on FDCs has been described as a prerequisite for the interaction of FDCs and B cells within lymphoid tissues. We found a strong up-regulation of ICAM-I after secondary stimulation with TNF-α. However, we already obtained functionally competent FDC-like cells capable of B-cell rosetting and antigen retention which strongly expressed CD35 under the TNF-α-free basic condition.
Development of functional germinal centres as well as the maintenance of FDC function has been claimed to strongly depend on the engagement of TNF family members (TNF-α and LT-α/β) [36
]. In-vitro generation of FDC-like cells, as described herein, appears to be widely independent of LT-α/β, whereas TNF-α may cause a further maturation of the FDC phenotype. The expression of HLA-DR which is controversial, but positively found by some authors [35
], was up-regulated by TNF-α as well (Table ).
Other factors were, however, inhibitory. Lymphocytes had to be eliminated as far as possible – which, especially when activated, inhibited the observed transdifferentiation. Specifically, IFN-γ may be the main factor: when added to monocytes as early as at the onset of culture, it strictly inhibited APpos
phenotype, as we have shown before [27
]. When added in a secondary step after 12 days of culture, IFN-γ up-regulated expression of CD54, but down-regulated CD35.
Of note, emperipolesis specifying the prolonged engulfment of B cells is a unique FDC property which, within germinal centres, is implicated in the protection of transiently internalised B cells from apoptosis [33
]. It is envisioned that this property may be useful for preserving immunogenic or tolerogenic B cells triggered in vitro
for their subsequent clinical application.
Identifying these newly differentiated cells as FDC-like cells culminates in the demonstration of storing antigen for long periods of time in its native form. This ability has been described as a unique feature of FDCs [34
] but, to our knowledge, neither for other types of dendritic cells nor for macrophages. Specifically, in FDC-like cells the antigen, provided as HRPO-IgG complexes, was retained in an enzymatic active form up to 16 days tested. Enzyme activity was localized perinuclearly. In contrast, macrophages eliminated the material within a few hours. Further studies will reveal the precise kinetics of uptake, sub-cellular distribution, storage and possible re-expression on the FDC surface by ultrastructural investigations.
Monocyte-derived FDC-like cells can now be generated in vitro
. Obviously, one pertinent employment of this system is to spur on the further elucidation of the underpinnings of immunological B-cell memory. Furthermore, cultures of FDCs may greatly facilitate research on the mechanisms underlying the capture and functional preservation of antigen. Next, because antigen retention is known to be exploited by pathogenic entities such as the human immunodeficiency virus-1 (HIV-1) [44
] as well as prions [45
], close investigation of these processes in vitro
provides hope for important progress in these areas. For example, Smith et al.
nicely demonstrated on isolated human FDCs, and in murine models alike, that FDC-entrapped HIV-1 remains highly infective for prolonged times, which appears to be due to protection of this virus from degradation [46
]. Last but not least, this novel protocol might even foreshadow later clinical applications of autologous FDCs or FDC-preserved B cells akin to the clinical studies currently conducted with T cell-tropic DCs. In conclusion, the present findings may open up new horizons for unraveling the secrets of antigen and pathogen storage by FDCs, with self-evident immunological and clinical implications.
Other groups have shown other cell types to transdifferentiate from monocytic precursors. However, all of these studies suffered from the lack of defined conditions (such as by using serum, conditioned media or undefined lymphocyte-derived signals) and/or employed a monocyte subfraction as the starter population. In contrast, we now introduce a common pathway, and define the signals required, for allowing monocytes to quantitatively convert into a state of high developmental plasticity from which FDC-like cells can be deduced. It needs to be stressed that our results do not per se
give a definitive clue as to the normal pathway of FDC differentiation. In fact, contemplating the actual existence of such a principal differentiation path may either unravel the main pathway of FDC generation or a salvage route, respectively. Normal FDCs in situ
express some myeloid markers such as CD14 and CD11b but not CD68. However, FDC tumors have been described to also express CD68 [47
]. In any case, the myeloid pathway desribed here obviously complements our previous findings on the ontogeny of T cell-directed DCs in that monocytes [20
] and their myeloid precursors [50
] can both be induced to quantitatively develop into T cell-tropic DCs, as is now commonly acknowledged.
Interestingly, AP and CD35 expression as well as antigen capture were occasionally observed to spontaneously emerge in single cells within control macrophage cultures. It thus becomes increasingly obvious that monocytes, macrophages, T-cell tropic DCs and FDCs belong to a continuum of a developmentally plastic system of cells which are highly interconvertible. Further data expand this view to osteoblasts, which were recently differentiated from a monocyte subset [24
], thus supporting our results on overlapping phenotypes between osteoclasts, osteoblasts, dendritic cells and macrophages [27
Moreover, as mentioned before, even cell classes generally considered to be developmentally distant from the myeloid lineage can be readily obtained from monocytes when the appropriate signals are provided. Hence, when acknowledging the monocyte as a previously ignored species of somatic stem cell, it appears unlikely that this fascinating cell type has already disclosed its entire potential. It will indeed be exciting to witness which ontogenetic surprises and therapeutic promises the future may hold.