Based largely on practical constraints, the insect immune and circulatory systems have been conceptually divided into discrete elements, and the immune system further dissected into cellular and humoral components 
. However, these entities are physiologically interrelated and have apparently evolved in integral association since the beginning of animal evolution 
. The cellular immune response remains only partially understood in mosquitoes as well as in other adult insects 
. Likewise, interactions between major circulatory elements and immune cells are virtually unknown and have received little attention. Using methods we previously developed for the study of hemolymph circulation 
, along with novel techniques for the in vivo investigation of hemocyte biology, we analyzed the cellular immune response in the pericardial region of the malaria mosquito, Anopheles gambiae
. We discovered that during an active infection, hemocytes migrate to the periostial regions, where they form a major component of the cellular immune response. Exemplifying the interrelationship of cellular immunity and circulatory processes, periostial hemocytes form phagocytic foci in regions of high hemolymph flow, which are also in the direct vicinity of the mosquito's major nephrocytes, the pericardial cells. Together, the data presented herein describe the formation of a novel immune tissue in mosquitoes. Because previous mosquito studies did not recognize sessile hemocyte aggregations as a major player in immunity, they failed to examine a large proportion of the cellular immune response, and thus, underestimated the relative contribution of hemocytes in anti-pathogen responses.
Using a correlative imaging approach, we scrutinized what appeared to be major phagocytic foci forming in the periostial regions of the mosquito heart. We previously reported that these foci form in response to infection, but their cellular composition and their functional role remained unknown 
. Similar foci form in Drosophila
, although they have never been directly studied 
. Here, we dispel the notion that phagocytosis on the surface of the heart is due to the activity of PCs. This finding was not entirely surprising, as in two different insect orders the PCs have been shown to be surrounded by a basement membrane 
, and the presence of this physical barrier should impede the direct phagocytosis of invading pathogens. Instead, we report that immune foci on the surface of the heart are composed of periostial hemocytes that rapidly and efficiently phagocytose pathogens. Rapid phagocytosis is believed to be an essential immune process, which culminates in pathogen death and the production of humoral immune components 
. The large number of periostial hemocytes present in infected mosquitoes, when compared to the total number of circulating hemocytes 
, suggests that this response involves between 10% and 25% of the hemocytes present in mosquitoes, thus highlighting the importance of pathogen sequestration in areas of high hemolymph flow. Finally, because nodulation and cellular encapsulation do not occur in mosquitoes 
, periostial foci formation is the primary hemocyte aggregation immune response in the culicid lineage.
Periostial immune aggregates are composed of a mixture of resident hemocytes and circulating hemocytes that settle in the pericardial regions in response to infection. Given that bacterial infection induces an increase in the number of circulating hemocytes in An. gambiae
, and that Aedes aegypti
hemocytes can replicate in response to various stimuli 
, we hypothesize that some of the migrating hemocytes seen in this study are the products of circulating hemocyte replication in response to infection. While the origin of periostial hemocytes seems clear, the molecular trigger that induces hemocyte aggregation in the periostial regions is unknown. The finding that hemocyte recruitment is induced by bacteria, Plasmodium
, carboxylate modified latex microspheres and soluble immune elicitors suggests that multiple pathways of immune activation can induce periostial hemocyte aggregation. Studies in other insects have identified several molecular components involved in sessile hemocyte aggregation and release. For example, Noduler mediates hemocyte aggregation in Lepidoptera 
, and multiple pathways mediate hemocyte proliferation and adhesion in Drosophila
. However, a great amount of genomic divergence is seen in insect immune genes 
, and thus, alternate pathways may be involved in mosquitoes.
What is perhaps most interesting about periostial hemocyte aggregates is their location. In dipterans, the cardiac ostia are the major incurrent valves for hemolymph entry into the heart 
. Thus, hemocyte aggregation in these regions greatly increases the likelihood that hemocytes encounter circulating pathogens, and that toxic products produced during pathogen breakdown are either immediately diluted as they are swept into the rapidly flowing hemolymph or are captured by the nephrocytic PCs that flank the heart. PCs filter proteins and colloids from the hemolymph 
, and thus, it is likely that the proximity of the PCs to the periostial hemocytes is essential for the quick absorption of pathogen breakdown products. We speculate that, during the course of arthropod evolution, periostial hemocytes and PCs adapted to their current locations because of their proximity to each other in an area of high hemolymph flow.
Data from the first hour post-infection showed that circulating hemocytes bind the alary or cardiac musculature within 100 µm of the ostia and then glide into the periostial regions at velocities that are orders of magnitude slower than that of the surrounding hemolymph flow. The molecular mechanism for this process was not a focus of this study, but the movement of insect hemocytes is known to be controlled by a number of different molecular pathways, and to be governed by processes such as adhesive capture and chemotaxis 
. The process observed here is likely a variant of adhesive capture, a process described in Drosophila
where injury induces the capture of hemocytes at epidermal wound sites 
. In agreement with the process of adhesive capture is our observation that mosquito hemocytes that reach the periostial regions originate from the circulating pool of hemocytes; video analyses show the binding of hemocytes to the alary and cardiac musculature and not the gliding of sessile hemocytes across extended distances. However, in contrast with adhesive capture in Drosophila
, where individual hemocytes arrive at injury sites and do not disperse or migrate 
, mosquito hemocytes bind within the general vicinity of the ostia and then move into their final point of attachment at velocities considerably slower than that of hemolymph flow. This spatially directed gliding process is likely mediated by shear-flow dynamics, a process that has been shown to drive cell migration toward areas of high flow in phylogenetically diverse organisms 
. This process is also reminiscent of the early stages of vertebrate leukocyte extravasation in response to inflammation, where activated endothelial cells produce factors that capture circulating leukocytes, at which point the leukocytes roll in the direction of flow, and then undergo diapedesis 
. Finally, on rare occasions hemocytes arrive into the pericardial regions as small aggregates. Thus, there remains the possibility that hemocyte aggregation in circulation, or increases in hemocyte size post-infection 
, are physical mechanisms for an evolutionarily pragmatic response to infection.
In a manner comparable to what we describe in mosquitoes, the macrophages of many lower vertebrates aggregate in areas of high blood flow in response to infection 
. These aggregates assemble in the spleen and liver, where they concentrate the destruction and recycling of exogenous and endogenous material. Likewise, the tissue macrophages of higher vertebrates (e.g., stellate cells and Kupffer cells) are believed to have originated from the same cellular ancestors as the phagocytes found in all animals, and are also commonly located in areas of high blood flow 
. It seems parsimonious to speculate that such phylogenetically disparate phagocyte aggregation responses are entirely based on functional analogy. However, mounting evidence suggests that these aggregation responses rely on conserved molecular and physiological components that were present in an ancient bilaterian 
. From a physiological perspective, we believe the data presented here exemplifies the taxonomically widespread importance of evolutionary constraints imposed on the cellular branch of the immune system as a consequence of its long history of evolution in close association with the circulatory system.
Although it is known that most Plasmodium
sporozoites rapidly die during their migration through the mosquito hemocoel 
, the specific interactions between sporozoites and hemocytes remain largely unknown. Earlier reports showed that phagocytosis of Plasmodium
by hemocytes occasionally occurs 
. Our data show the in vivo interaction between Plasmodium
and hemocytes, along with the first evidence of a systemic cellular immune response to late-stage malaria infection (increase in periostial hemocyte numbers). While the large number of sporozoites released by each oocyst makes it unlikely that phagocytosis is the primary component of the anti-Plasmodium
response in the hemocoel, increases in melanization and periostial hemocyte aggregation suggest that hemocyte activation leads to the production of humoral factors that target Plasmodium
via lytic and melanization pathways. Evidence from others supports this idea, as Plasmodium
development in mosquitoes induces the transcriptional regulation of immune genes in hemocytes 
, and our data on melanin deposition near hemocyte-Plasmodium
interactions are in agreement with studies on the anti-sporozoite response 
In most insects, one problem foiling hemocyte research is that no effective means of specifically staining hemocytes in vivo exists 
. As part of this investigation, we developed a CM-DiI based method that fluorescently labels hemocytes in vivo. CM-DiI is a lipophilic probe with high affinity to plasma membranes. While this probe stains all cells grown in culture, it only stains hemocytes when injected into mosquitoes. Several lines of evidence support the specificity and efficiency of CM-DiI hemocyte labeling. First, CM-DiI stains virtually all circulating hemocytes and also stains cells attached to tissues in a random pattern (sessile hemocytes). Second, CM-DiI stains cells that fit the morphological description of hemocytes given by previous authors 
. Finally, the vast majority of cells that stain with CM-DiI exhibit the characteristic phagocytic signature of mosquito granulocytes. Although the mechanism by which CM-DiI specifically stains hemocytes remains unknown, our data suggest that CM-DiI is unable to cross the basal lamina surrounding internal tissues, and that it initially stains hemocytes by binding their membranes and then becoming subsequently phagocytosed. Evidence supporting this mechanism of labeling includes: (1) co-injection of formaldehyde along with CM-DiI eliminates hemocyte specificity and all tissues become labeled; (2) incubation of CM-DiI injected mosquitoes in a solution containing a detergent releases the hemocyte-captured CM-DiI and leads to the staining of other tissues; (3) injection of carbon particles prior to CM-DiI treatment blocks hemocyte staining; (4) CM-DiI staining appears most brightly as puncta within the hemocytes but over time spreads over the entire cell membrane; and (5) within minutes of mixture with PBS, CM-DiI precipitates out of solution and loses its hemocyte staining efficacy. Taken altogether, this suggests that CM-DiI could be categorized as a functional marker that stains only hemocoelic phagocytes in vivo. Techniques based on similar principles have been used for the study of macrophage biology in mammals 
. Given the technical and practical difficulties associated with the creation of transgenic mosquito strains 
, as well as the fact that some of the more common Drosophila
hemocyte markers are not encoded in the mosquito genome (e.g., hemese), the CM-DiI approach described here for the first time allows the study of mosquito hemocyte cell biology in vivo and in real time. We expect that this procedure could be adapted for the study of hemocyte biology in a broad range of insects.
In conclusion, there remain deficits in our current knowledge of hemocyte biology in adult insects, as well as in our understanding of the direct interactions between the insect circulatory and immune systems. Here, we developed new methods for the in vivo study of mosquito hemocytes and pericardial cells (nephrocytes), and applied these methods to discover a novel mosquito immune response. Namely, we uncovered periostial hemocyte aggregates, an immune tissue that is located on the surface of the mosquito heart and represents a basal component of the cellular immune response against bacteria and malaria parasites.