Blood cell progenitor maintenance: Collier barks out of the niche

ABSTRACT

Drosophila lymph gland, a larval haematopoietic organ, has emerged as a popular model to study regulatory mechanisms controlling blood cell progenitor fate. In this organ, the Posterior Signaling Center (PSC), a small group of cells expressing the EBF transcription factor Collier, has been proposed to act as a niche required for progenitor maintenance. Accordingly, several reports showed that PSC size/activity modulation impacts on blood cell differentiation. Yet our recent results challenge this model. Indeed, we found that PSC ablation does not affect haematopoietic progenitor maintenance. This unexpected result led us to reinvestigate the role of the PSC and collier in hematopoiesis. Consistent with previous findings, the PSC appears required for the production of a specialized blood cell type in response to parasitization. Moreover, our results indicate that the massive blood cell differentiation observed in collier mutant larvae is not due to the lack of PSC but to collier expression within the haematopoietic progenitors. We thus propose a paradigm shift whereby larval blood cell progenitor maintenance is largely independent of the PSC but requires the cell-autonomous function of collier.

Drosophila blood cells

The balance between haematopoietic progenitor maintenance and differentiation into specialized blood cell types is a finely tuned process.¹ Indeed, the timely and scaled production of different blood cell lineages is required for proper development as well as for ensuring tissue homeostasis or mounting a proper response to infections. Thanks to the phylogenetic conservation of several aspects of blood cell development, Drosophila melanogaster has emerged as a fruitful model to identify genes regulating blood cell fate and to assess the relationships between haematopoietic progenitors and their (micro)environment during normal development or in response to pathological situations.^2-4

There are two waves of generation of blood cell progenitors (prohemocytes) during Drosophila development: first in the embryo and second in the larval lymph gland.^5,6 These prohemocytes can differentiate into three mature blood cell types (hemocytes) that are functionally related to mammalian myeloid cells: plasmatocytes, crystal cells and lamellocytes, implicated in phagocytosis, melanization and encapsulation, respectively. Plasmatocytes and crystal cells are observed in all stages of fly development whereas lamellocytes are produced only during larval life following a response to some immune challenges such as wasp parasitism. Of note, lamellocytes and crystal cells can also form by transdifferentiation of mature plasmatocytes.^7-10 These embryo-derived blood cells are found in the larvae either as patches of cells attached to the cuticle (sessile hemocytes) or circulating in the hemocoel and persist until adulthood.^7,11 Blood cells that are formed during the second wave of hematopoiesis, in the larval lymph gland, are normally released into circulation only at pupariation as the lymph gland bursts, and they also populate the adult fly.^12-14 While it was long believed that adult Drosophila solely relies on the pool of differentiated hemocytes formed during the embryonic and larval waves of hematopoiesis, a recent report indicates that blood cell proliferation and differentiation also takes place in abdominal patches located close to the heart during adulthood.¹⁵

The lymph gland and the niche

Over the last few years, a number of studies have assessed the mechanisms controlling blood cell development and homeostasis in the lymph gland. Indeed, the organization of this organ makes it particularly suitable to analyze the balance between blood cell progenitor maintenance and differentiation.¹³ In the third instar larvae, the lymph gland is composed of a pair of anterior/primary lobes and several pair of small posterior lobes (Fig. 1A,B).¹² The posterior lobes mostly consist of blood cell progenitors whereas the primary lobes contain blood cell progenitors in their medullary zone (MZ) and a mixture of maturing and fully differentiated hemocytes in their cortical zone (CZ).^13,16 In addition, a small group of ±50 non-haematopoietic cells at the posterior of each primary lobe form a structure called the posterior signaling center (PSC).¹⁷

Figure 1.

(A) Schematic representation of the lymph gland in third instar larvae. The primary lobes are subdivided in three domains: the cortical zone (CZ, red) contains differentiated hemocytes that stemmed from the medullary zone (MZ, blue), which is composed of progenitors, and the posterior signaling center (PSC, green). The posterior lobes (2 to 4 pairs) contain mostly blood cell progenitors (blue). The green arrows schematically represent the presumptive PSC-dependant prohemocyte maintenance signals and the yellow arrows indicate the main axes of differentiation. (B) Confocal section showing the PSC (green), blood cell progenitors (blue) and differentiated hemocytes (red) in the anterior lobes of a third instar lymph gland. (C, D) Expression of the prohemocytes marker tepIV (red) and the PSC marker col-GAL4,UAS-mCD8GFP (col>GFP, green) in control third instar lymph gland (col>GFP) or upon Myc overexpression in the PSC (col>GFP>myc). Myc overexpression causes an increase in col>GFP⁺ cell number but also the ectopic expression of tepIV in these cells. Dashed white lane: primary lobes contour. (C'-C”’, D'-D”’): high magnification views of the PSC region; merge channel (C', D') and separate channels (C,“C”’, D, “D”’).

In 2007, two seminal studies suggested that the PSC acts as a haematopoietic niche,^18,19 i.e. a specialized microenvironment, that provides signaling molecules required for maintenance of blood cell progenitors.²⁰ These studies revealed an interesting parallel with the Haematopoietic Stem Cell (HSC) niche in mammalian bone marrow, and several publications lent support to this hypothesis.^5,21 Below, we summarize the main evidence for this model.

One key argument stems from the analysis of mutant larvae lacking a PSC. In particular, it was shown that PSC development depends on the EBF transcription factor Collier (Col) and the HOX factor Antennapedia (Antp).^18,19,22 In the embryo, Col expression initially covers the whole primary lobe precursors before being restricted to the posterior part of the lobes where, together with Antp, it prefigures the PSC. In col mutant, the PSC is not specified and larval lymph glands exhibit a massive decrease in blood cell progenitors and a concomitant increase in hemocyte differentiation both in anterior and posterior lobes.^18,19 Thus, it was concluded that the PSC is absolutely required for progenitor maintenance. Similarly, it was reported that Antp mutant larvae had no PSC and displayed an increase in crystal cell differentiation.¹⁹ In that case though, the status of the progenitor population was not assessed.

Another important piece of evidence for the PSC/niche hypothesis came from the analysis of its signaling activity. In particular, the PSC expresses Hedgehog (Hh) and, using a thermosensitive allele, it was shown that hh inhibition leads to increased plasmatocyte and crystal cell differentiation.¹⁹ Importantly, it was found that Hh and JAK/STAT signaling pathways are active in the MZ and interfering with theses pathways resulted in increased differentiation in the primary lobes.^18,19 These data suggested that the PSC acts in a non-cell autonomous manner to activate Hh and JAK/STAT signaling in prohemocytes and thereby preventing their differentiation. Interestingly, PSC cells extend long filopodia toward the MZ that might directly deliver pro-maintenance signals to the prohemocytes.^18,19,23 Yet the functional importance of these cytoplasmic extensions has not been demonstrated.

Finally, genetic manipulations of PSC size also lend support to the PSC/niche model. In particular, it was shown that Antp over-expression in the PSC leads to an expansion of the PSC and of the MZ, suggesting that an increase in signaling from the PSC promotes progenitor maintenance.¹⁹ Similarly, some subsequent studies that identified genes controlling PSC cell number also found a correlation between PSC size and hemocyte differentiation.^23-29 For instance, decreased PSC size caused by inhibition of Wg signaling or Arf1 expression in the PSC resulted in increased hemocyte differentiation.^24,29 Conversely, increased PSC size caused by InR/TOR pathway over-activation or decreased BMP signaling promoted progenitor maintenance at the expanse of hemocyte differentiation.^27,28 Moreover, decreasing the expression of the PDGF- and VEGF-receptor related (PVR) ligand Pvf1 by the PSC also impaired prohemocyte maintenance.²⁵ In that case, PVR signaling acts on differentiated hemocytes that secondarily promote progenitor maintenance via the Adenosine deaminase-related growth factor A (ADGF-A) pathway.

All together, these data are consistent with the idea that the PSC operates as a niche expressing different signaling molecules that promote directly or indirectly blood cell progenitor fate maintenance.

Revisiting the role of the PSC

Yet, some morphological features of the lymph gland are at odds with this model. First, the topology of the primary lobes is not fully consistent with the hypothesis that the PSC is a source of pro-maintenance signals: differentiated hemocytes are often observed in the PSC vicinity whereas some blood cell progenitors are very distant from it (Fig. 1B). Actually, the global differentiation axis of the primary lobes is rather perpendicular to the heart tube than originating from the PSC (Fig. 1A). This unexpected organization might be explained by feedback loops from the mature hemocytes toward the progenitors or by some physical constraints and it would be interesting to try to model primary lobes development to gain insights into their patterning. Second, it is unclear how the posterior lobes, that are almost exclusively composed of blood cell progenitors in third instar larvae and lack a PSC structure, can be regulated by this niche: PSC filopodia are oriented toward the MZ of the primary lobes, and PSC cells are physically separated from the secondary lobes by pericardial cells and even further away from tertiary or quaternary lobes. Most studies focused their analysis on primary lobes and our knowledge of posterior lobe development and regulation remains very limited. Their development is heterochronic: they appear and mature later than the primary lobes.³⁰ The “younger” status of these prohemocytes might allow them to function without PSC pro-maintenance signals. However, in col mutant larvae, they massively differentiate.¹⁸

These and other unpublished observations led us to test what happens if we remove the PSC. To do so, we expressed the proapopotic gene reaper (rpr) in the PSC.³¹ This strategy allowed us to efficiently kill PSC cells at least from the 2^nd instar larval stage onwards and to recover PSC-less lymph glands. According to the prevailing model, one would have expected that most blood cell progenitors differentiate in this condition. In contrast, quantitative analyses revealed that PSC-less lymph glands harbored a normal proportion of blood cell progenitors and did not exhibit increased blood cell differentiation. To circumvent possible drawbacks associated with Rpr expression, we also re-analyzed the phenotypes of col and Antp mutant larvae. While both genes are required for PSC formation, their mutation led to different phenotypes: blood cell progenitors were still present in Antp mutants whereas they were almost completely absent in col larvae. Thereby, it appears that the PSC is largely dispensable for lymph gland homeostasis and in particular for prohemocyte maintenance.³¹

One important difference between our strategy and previously published one is that we tried to set the niche activity to zero rather than modulate it by increasing/decreasing its size or the expression of signaling molecules. It is clear that the later strategy helped pinpoint the genes controlling PSC growth and showed that signaling from the PSC can influence blood cell progenitor maintenance. However, the interpretation of these results is not always straightforward. For instance, it was proposed that the increase in PSC size caused by col down-regulation does not affect hemocyte differentiation because it is compensated by a drop in Hh expression.²⁸ Also modulation of PSC size/activity might lead to the activation of inappropriate gene expression programs. For instance, Myc overexpression causes a ∼5 folds PSC cell number increase but now these cells also illegitimately express the prohemocyte marker tepIV, suggesting that they might have changed fate.(Fig. 1D)

Of note, one line of evidence that supported the PSC niche function is that it expresses Hh, and that inhibition of this signaling pathway in the prohemocytes promotes their differentiation.¹⁹ Since blood cell progenitors are maintained in PSC-less larvae, it will be interesting to study the presence of other sources of Hh than the PSC. Interestingly, some scattered Hh positive cells have been observed within the Cortical Zone.^19,23 This observation is reminiscent of Serrate expression. Initially, the PSC was identified as a group of cells expressing this Notch ligand and proposed to act as a signaling center that induces crystal cell differentiation by activating the Notch pathway.¹⁷ However, as crystal cells still differentiate in col mutant lymph glands, it was proposed that dispersed cells also expressing Serrate outside the PSC might be the ones responsible for crystal cell induction.²² Accordingly, these so-called Lineage-Specifying Cells have recently been shown to trigger crystal cell differentiation.³² Along the same line, it will be interesting to characterize the population of Hh expressing cells present within the cortical zone and to study their impact on prohemocyte maintenance.

Another role that was assigned to the PSC is to control lamellocyte differentiation. Indeed, lamellocytes fail to differentiate in col mutant larvae upon infection by eggs from the parasitoid wasp Leptopilina boulardi.²² Furthermore, parasitization induces an increase in NF-κB signaling and ROS levels in the PSC that leads to lamellocyte differentiation.^26,33 Accordingly, previously published and our own PSC ablation experiments indicate that PSC-less larvae do not produce lammellocytes in response to wasp infection.^18,31 Thus, several lines of evidence indicate that the PSC is required to mount a proper cellular immune response following parasitization. It will be of particular interest now to identify the initial systemic signal that alters PSC signaling and causes lamellocyte production upon wasp infestation.

In summary, our analysis of PSC-less larvae (either in Antp mutant or following rpr expression) demonstrates that this structure is not absolutely required for blood cell progenitor maintenance under normal breeding conditions and undermines the notion that the PSC is the lymph gland haematopoietic niche.^31,34 In view of the current literature, it is tempting to speculate that PSC activity is more complex than just sending pro-maintenance signals and that it could also be a source for pro-differentiation molecules. This is consistent with its role during wasp infestation where it acts as a source of lamellocyte-inducing factors, and it will be worth testing whether it is involved in other stress responses. Moreover, in mammals the haematopoietic niche is a complex structure composed of several cell types, including osteoblast, perivascular cells, endothelial, Schwann cells, sympathetic neuronal cells and macrophages.³⁵ The Drosophila lymph gland haematopoietic niche is probably a complex system too: differentiated blood cells participate in blood cell progenitor maintenance,²⁵ and other components of the niche are likely to be identified in the future.

A key cell-autonomous role of collier for haematopoietic progenitor maintenance

Clearly, PSC-ablated or Antp mutant lymph glands don't phenocopy the massive differentiation of blood cell progenitors observed in col mutants.³¹ The difference between these three kinds of PSC-less larvae suggested that col is required outside of the PSC for prohemocyte maintenance. Indeed, beside its high expression in the PSC, col is expressed at low level in the prohemocytes not only within the medullary zone of the primary lobes but also in the posterior lobes. Importantly, inhibiting col expression in the prohemocytes is sufficient to cause their differentiation even in the presence of PSC.³¹ Conversely, Col overexpression in prohemocytes impairs their differentiation,^18,31 and re-expressing Col in these cells is sufficient to ensure their maintenance in a col mutant background.³¹ Thus it appears that prohemocyte differentiation in col mutant lymph gland is due to the lack of Col expression within the prohemocytes and not to the absence of PSC.

These results raise several questions concerning the regulation of col and its mode of action. Intriguingly, col is expressed at high level in the PSC but also in some cells of the posterior lobes.³¹ The col-GAL4 transgene classically used to label the PSC mirrors the domains of high expression but fails to drive where col is expressed at low level.^18,31 The cis-regulatory module(s) mediating col low expression and the signaling pathways/ transcription factors controlling its activity remain unknown. Within the primary lobes, col is expressed in a sub-domain of the medullary zone (as labeled by the expression of the JAK/STAT receptor Domeless) that overlaps with tepIV expression and lies closer to the heart tube.³¹ col / tepIV expression may thus label a more primitive population of prohemocytes. Different studies have shown that lymph gland organization is more complex than the basic MZ/CZ zonation and that prohemocytes are heterogenous.^5,13,30,36 It will be interesting to better address the complexity of the lymph gland blood cell progenitors and to reveal their true potential. In addition, further experiments will be required to assess the epistatic relationships between col and other known regulators of prohemocyte fate. For instance, col was shown to be required downstream of JAK for JAK/STAT pathway activation in cell culture.³⁷ Whether it plays a similar role in the lymph gland remains to be shown. Also, it will be worth characterizing the genes regulated by Col in the prohemocytes that could confer a blood cell progenitor state. Actually, direct target genes of Col identified in the larval sensory neurons include tepIV as well as several well-known regulators of larval blood cell development.³⁸ Another outstanding question will be to decipher the regulatory code of Col in the prohemocytes versus the PSC: does it depend on Col level, on the presence of cell-type specific co-factors or post-translational modifications?

In conclusion, our study leads to a paradigm shift concerning Drosophila lymph gland hematopoiesis: the PSC is not the niche required for blood cell progenitor maintenance. Keeping with its name, it primarily acts as a signaling center that can fine-tune the balance between progenitor maintenance and differentiation and that is required for lamellocyte differentiation following wasp infestation. Future experiments aiming at deciphering the role of the PSC in other stress situation may bring further insights into its regulatory function and on the (micro)environmental control of blood cell homeostasis. Moreover, we revealed the essential, cell autonomous function of the EBF transcription factor Col in haematopoietic progenitor maintenance. This opens interesting avenues of research concerning the putative role of its homolog in mammalian haematopoietic stem cells. It is anticipated that analyses of Col function in the PSC and prohemocytes will bring valuable insights on the molecular mechanisms of regulation of blood cell development.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.