|Home | About | Journals | Submit | Contact Us | Français|
Tumor immunosurveillance is a well-established mechanism for regulation of tumor growth. In this regard, most studies have focused on the role of T- and NK-cells as the critical immune effector cells. However, macrophages play a major role in recognition and clearance of foreign, aged, and damaged cells. Macrophage phagocytosis is negatively regulated via the receptor SIRPα upon binding to CD47, a ubiquitously expressed protein. We recently showed that CD47 is up-regulated in myeloid leukemia and migrating hematopoietic progenitors, and that the level of protein expression correlates with the ability to evade phagocytosis. These results implicate macrophages in immunosurveillance of hematopoietic cells and leukemias. The ability of macrophages to phagocytose tumor cells might be exploited therapeutically by blocking the CD47-SIRPα interaction.
Over the last 30 years, the dominant paradigm in cancer biology has held that tumor cells successively acquire multiple mutations and inherited changes in gene expression by epigenesis over a lifetime that allow cells to progress from normal tissue, to pre-malignant dysplastic growths, and eventually to fully malignant tumors. These mutations or epigenetic changes have been considered to be of two general types: some result in over-activity of genes that promote cancer growth (oncogenes), while others result in inactivation of genes that restrict tumor growth (tumor suppressors). Identification of mutations acquired in cancer, and determining their role in tumor progression, has been the fruitful result of the last few decades of research.
Perhaps the most informative cancer progression model has been the colorectal cancer model, described as a multistep set of events by Fearon and Vogelstein1. Hanahan and Weinberg famously distilled the knowledge gained from many studies to enumerate six hallmarks that most cancers share: self-sufficiency in growth signals, insensitivity to growth-inhibitory signals, evasion of programmed cell death, limitless replicative potential, sustained angiogenesis, and tissue invasion and metastasis2. While this is a simplification of a complex process, these capabilities are almost universally found in varying degrees in tumors, indicating their profound importance in neoplastic progression. Nonetheless, recent work has shown that two more hallmarks may also be essential for development of cancer, namely the ability of cancer cells to acquire stem-cell like machinery to allow self-renewal (or for the inciting mutations to arise in a bona fide tissue stem cell) 3, and the ability of cancer cells to evade elimination by the immune system. The cancer stem cell model has been reviewed extensively elsewhere4. Here we will instead focus on the role of macrophages as potential mediators of immunosurveillance, and its clinical implications.
The “tumor surveillance” hypothesis was first proposed by Macfarlane Burnet and Lewis Thomas in 19575. The hypothesis suggests that tumor cells might have “altered-self” antigens that allow the immune system to recognize them leading to the tumor’s subsequent eradication. Recently, researchers have created mice with genetic deficiencies in various immune functions. For instance, mice that lack genes essential for immune development or effector responses, such as recombination activation gene 2 (RAG2−/−), interferon gamma (IFNγ−/−), or perforin (Pfp−/−), among others, have increased incidences of spontaneous and 3-methylcholanthene induced cutaneous tumors and fibrosarcoma as well as increased incidence of spontaneous lymphomas and colon, breast, and lung carcinomas, indicating that a functional immune system is needed for tumor surveillance of at least some malignancies6,7. Furthermore, in the last 10 years, hundreds of antigens expressed by human tumors have been identified that have limited or no expression by the vast majority of cells in normal tissues8,9, allowing for the possibility of the adaptive immune system to recognize tumors as foreign.
While the adaptive immune response is now well recognized to be an important component of anti-tumor immunity, the role of the innate immune system has begun to be elucidated as well. Since the seminal discovery that NK cells can eliminate endogenous cells with decreased class I MHC expression10, the idea was born that tumors could be distinguished from normal tissues because of “missing self”. Since these landmark findings were first published, a wealth of information about activating and inhibitory receptor-ligand interactions on NK cells has been discovered. The most interesting of these, from a tumor immunology perspective, has been the natural killer group 2, member D (NKG2D) receptor, which can recognize MHC class I related proteins A and B (MICA and MICB)11, and can also recognize antigens ectopically expressed on tumors, such as Rae1 or H6012, leading to the destruction of tumor cells which over-express these proteins. Furthermore, it has recently been discovered that DNA damage can itself lead to upregulation of NKG2D ligands13, thereby linking pro-tumorigenic insults with the markings of “altered-self”.
While there has been much data to support a role for T- and NK-cells in tumor surveillance, to date there has been a dearth of evidence to indicate that macrophages can directly kill tumor cells via phagocytosis. In the 1980s and 90s, there was interest in the idea that Kupffer cells, specialized macrophages that reside in the liver sinusoids, could prevent metastases from engrafting in the liver, but this avenue of research has been largely abandoned14-18. More recently, investigators have focused on the ability of macrophages to infiltrate solid tumors and modulate T-cell and stromal activity to either favor or inhibit tumor growth19. The general consensus is that macrophage activity in such situations is pro-tumorigenic via their ability to promote angiogenesis and metastases20. These so-called tumor associated macrophages (TAMs) are thought to be recruited by tumor derived chemokines and converted to a tumor permissive state21. These TAMs are called M2, or alternatively-activated macrophages, to distinguish them from M1, or classically-activated macrophages, and have been extensively studied22-24. Alternatively-activated macrophages produce less inflammatory cytokines, promote Th2 T-cell responses, favor wound healing, and increase local angiogenesis25. Whether these TAMs can be converted from the pro-tumorigenic M2 type to the pro-inflammatory M1 type to attack tumors is still under investigation.
Perhaps one reason for the lack of evidence for the involvement of phagocytosis in tumor killing is that, until recently, many of the molecules involved in macrophage recognition of cells and particles had not been determined. The evolutionarily oldest of these recognition receptors are the C-type lectin receptors, which mediate binding to carbohydrate moieties on bacteria26. In higher organisms, phagocytes can recognize foreign cells and particles that have been coated with antibody or complement via Fc receptors or complement receptors27 (i.e. opsonization). Scavenger receptors were originally identified by their ability to recognize modified low density lipoprotein, but are now known to recognize bacterial cell wall components such as lipopolysaccharide (LPS) as well28. In addition, the extracellular matrix proteins fibronectin and vitronectin can function as non-specific opsonins via integrin receptor mediated phagocytosis27. Apoptotic cells express the ligand phosphatidylserine (PS) on their cell surface29, and recognition by the PS receptor (most likely Tim430) on macrophages ensures prompt removal of apoptotic cells before they lyse and incite an inflammatory response. Even in the absence of apoptosis, neutrophils age over the day of their life, and when old, convert from being resistant to being sensitive to macrophage phagocytosis31; this process can regulate neutrophil numbers even when they have an artificially extended lifespan by enforced expression of Bcl2.
While these studies have illustrated the large variety of molecules involved in phagocytosis of foreign or apoptotic cells, it is the recent identification of ligands expressed on living self cells that can influence phagocytosis that might have more relevance for tumor surveillance. One such molecule is the pentaspanin CD47. Also known as integrin associated protein (IAP) because of its physical association with the integrins avb3, aiibb3, and a4b132, CD47 was first cloned as a marker expressed in ovarian tumors33. Since then, it has been shown to be ubiquitously expressed on mammalian tissues in a variety of splice isoforms34. The CD47−/− mouse has a generally mild phenotype with the only dramatic deficit being rapid death following experimental induction of bacterial peritonitis, presumably due to defects in neutrophil migration35. CD47 has also been implicated in cell migration36,37, T-cell co-stimulation38, neuronal development39 and thrombospondin binding40.
The first identification of a role for CD47 in phagocytosis was the observation that red blood cells (RBCs) from mice deficient in CD47 were rapidly cleared upon transfusion into wild-type hosts41. Indeed, as RBCs age, they lose expression of CD47 indicating that this may be a physiological mechanism by which they are cleared42. This observation was subsequently extended to leukocytes, where it was observed that CD47−/− bone marrow was unable to rescue wild-type mice after lethal irradiation43. The mechanism by which CD47 regulates phagocytosis is by interaction with signal regulatory protein alpha (SIRPα) on macrophages44-46. Upon ligation with CD47, the immunoreceptor tyrosine based inhibitory motif (ITIM) of SIRPα becomes phosphorylated and recruits SHP phosphatases to the membrane, resulting in inhibition of myosin accumulation at the cell surface, which thus precludes phagocytosis47. Thus, CD47 appears to be a required signal to prevent macrophage phagocytosis of self. More recently, studies have described phagocytosis as the result of a balance between pro- and anti-phagocytic ligands. For example, when calreticulin, a pro-phagocytic cell surface marker, is blocked, CD47 expression is no longer required to prevent default phagocytosis48. Seen in this light, any given macrophage encounter will result in a series of stimulatory and inhibitory inputs, the sum of which determines whether phagocytosis is triggered (Figure 1a).
It is important to note that while many interactions can positively regulate phagocytosis, SIRPα-CD47 is the only known receptor-ligand pair that is a negative regulator of macrophage phagocytosis, highlighting the importance of CD47 for prevention of self phagocytosis. Some authors have proposed that plasminogen activator inhibitor-1 (PAI-1) has a “don’t eat me” function based on evidence that PAI-1−/− or anti-PAI-1 antibody treated neutrophils are phagocytosed more readily than their wild-type counterparts49. Thus far, a receptor for PAI-1 on macrophages has not been identified. Its effect may instead be due to a cis inhibitory interaction with calreticulin. In the absence of PAI-1, calreticulin appears to be exposed on viable neutrophils, allowing increased clearance by macrophages. Another molecule, CD31, can negatively regulate tethering of viable leukocytes to macrophages, and becomes disabled once cells undergo apoptosis. However, the function of this molecule appears to be limited to tethering, and does not appear to directly affect the phagocytic machinery in macrophages50. A third molecule, CD200, has been shown to negatively regulate macrophage function in cancer. However, it functions primarily by altering inflammatory status and macrophage activation, and is not a phagocytic regulatory molecule per se (reviewed in 51). What these studies reveal is that phagocyte activation, binding, and engulfment is a complex, multi-step process with many players; however no ligand other than CD47 has been found that has a direct effect on phagocytosis via a macrophage inhibitory receptor. It remains to be seen if one of these signals is dominant over the others in the process of phagocytosis.
CD47 is ubiquitously expressed at a low level on all hematopoietic cells, and its species-specificity may underlie some aspects of hematopoietic xenotransplant rejection 52-54. Indeed, immunodeficient non-obese diabetic, IL-2 receptor common gamma chain deficient (NSG) mice express a SIRPα variant that is better able to interact with human CD47, which may explain the tolerance for human hematopoietic grafts in these mice relative to other immunodeficient strains54. Our recent evidence indicates that CD47 expression switches from low baseline levels to high levels rapidly on normal hematopoietic stem cells (HSCs) when they mobilize into the periphery after a strong inflammatory stimulus, such as granulocyte colony stimulating factor (GCSF) or (LPS)55. This upregulation is temporally precise, as CD47 levels quickly return to baseline within days. We hypothesize that this up-regulation is necessary to prevent clearance of normal, healthy hematopoietic cells both during their constitutive recirculation from marrow to blood to marrow, and also during inflammation, which also mobilizes HSC and progenitors; a low-level of CD47 at steady state is sufficient for prevention of phagocytosis. In support of this model, 10 to 20% of the normal red blood cell density of CD47 is sufficient to prevent clearance of polysterene beads by non-activated macrophages in the absence of other pro-phagocytic ligands47. Furthermore, our evidence indicates that genetically reducing levels of CD47 on normal murine stem and progenitor cells in vivo, or antibody-mediated blocking of CD47 of normal human HSCs in vitro, is not sufficient to induce phagocytosis by macrophages at steady state. Our data also show that HSCs that are haplosufficient for CD47 (i.e. half the usual expression level) are more likely to be cleared after administration of LPS, suggesting that CD47 upregulation may occur to protect migrating HSCs from being phagocytosed as they transit past activated macrophages into hematopoietic tissues during the acute inflammatory response (reviewed in 56). LPS has long been known to cause macrophage activation; one mechanism by which this may occur is through down-regulation of SIRPα expressed on some macrophage subsets after exposure to LPS57, which may lead to a decreased threshold for phagocytosis. However, by increasing CD47 expression levels, sufficient engagement of SIRPα may still occur (since this interaction may display high avidity characteristics like other immune receptor-ligand interactions58), hence those cells would not be phagocytosed during this activated state (Figure 1c-d). In addition, pro-phagocytic markers may also be increased after mobilization/inflammation (see below), making CD47 up-regulation even more critical for cell survival.
Leukemic stem and progenitor cells from chronic myeloid leukemia (CML) during blast crisis and acute myeloid leukemia (AML) also constitutively upregulate CD47 on their cell surface55,59. To test whether CD47 upregulation affects tumorigenicity, we engineered a human AML cell line that does not engraft in RAG2−/−, common gamma chain −/− mice (mice lacking T, B and NK cells, but that still have functional phagocytes), to express varying levels of murine CD47. The pathogenicity of the cell line in vivo, and the ability to evade phagocytosis in vitro, correlated with the level of CD47 expressed. Thus, the ability of leukemia cells to constitutively maintain high levels of CD47 is yet another example of a mechanism by which a tumor stem cell dysregulates a normal physiological process during pathogenesis.
Why would hematopoietic malignancies require additional protection against phagocytosis? One hypothesis is that the process of tumorigenesis itself creates a pro-inflammatory environment (Figure 2a). This could be due to the ability of normal stem cells to repopulate local or distant empty niches60-63, which is a migration begun by detaching from neighbor cells and could involve local dissolution of extracellular matrix to travel to adjacent or distant sites. Such an environment might lead to increased risk of phagocytosis due to increased macrophage activity and/or increased levels of opsonins, such as immunoglobulin in the serum or from dissolved components of the extracellular matrix such as fibronectin and vitronectin. In this case, there would be a strong selective pressure for those tumor clones that can escape phagocytosis by upregulating CD47.
A second hypothesis is that tumor cells themselves are more “phagogenic”, i.e. they are inherently more prone to eliciting phagocytosis, possibly because they express increased pro-phagocytic ligands, and escape variants eventually arise through selection that upregulate CD47 to compensate (Figure 1e-f, ,2b).2b). Similar to the way that NK cells patrol the epithelial environment for “stressed self”, macrophages may preemptively clear altered cells in the hematopoietic environment. Given that all major hematopoietic tissues are heavily infiltrated by macrophages, particularly at their sites of entry, phagocytosis of potentially dangerous variants could provide a potent tumor immunosurveillance mechanism. Moreover, identifying the pro-phagocytic ligands expressed on tumor cells could be useful in designing new therapies. Calreticulin is again implicated as one such molecule; it is upregulated in cells after irradiation64 or anthracycline administration65, providing a potential link between pro-tumorigenic insults and phagocytic clearance.
It remains possible that CD47 up-regulation has functions in addition to prevention of phagocytosis, as described above. In addition to its ability to bind to extracellular receptors, CD47 also has the ability to transduce signals into the cell after interaction with thrombospondin or integrins, leading to biological activities such as cell spreading, migration, or even apoptosis32. Our recent studies demonstrate that CD47 up-regulation does affect phagocytosis, though other functions may also be altered in these cells. However, the targeting of CD47 by therapeutic molecules in AML appears to primarily abrogate its function as a “don’t eat me signal”, without affecting survival or migration, leading us to speculate that its dominant function in vivo, in at least leukemic cells, is to prevent phagocytosis (Figure 1g).
We showed that upregulation of CD47 in human myeloid leukemias and their tumor-initiating leukemia stem cells contributes to pathogenesis through the engagement of CD47 with SIRPα leading to inhibition and evasion of phagocytosis by the host innate immune system55,59. This finding has therapeutic implications in that blockade of the CD47-SIRPα interaction by a blocking monoclonal antibody against CD47 could potentially enable phagocytosis of leukemia cells. Thus, we investigated whether antibody-mediated CD47-SIRPα blockade could enable phagocytosis of CD47-expressing leukemias59. In our studies, monoclonal blocking antibodies targeting CD47 preferentially enabled phagocytosis of Lin-CD34+CD38-CD90- human acute myeloid leukemia stem cells (AML LSC, rare cells within the neoplasm capable of recapitulating disease in a xenotransplant model) by macrophages in vitro. When administered in vivo to human AML engrafted mice, anti-CD47 antibody eliminated leukemic disease and tumorigenic potential through targeting of AML LSC. Although this antibody is effective in eliminating AML in mouse models, there is accumulating evidence that anti-CD47 antibodies may be effective in other tumors. Blocking antibodies directed against CD47 have been shown to eliminate human chronic lymphocytic leukemia B cells in vitro66, a multiple myeloma cell line67 and lymphoid leukemia cell lines68 in vivo. In solid tumors, anti-CD47 antibodies have been shown to eliminate human head and neck squamous cell carcinoma cell lines69 as well as primary human bladder cancers in vitro70. Such reports demonstrate that antibody-mediated blockade of CD47 signaling may be therapeutic in multiple hematopoietic and solid tumor malignancies.
What is the mechanism behind anti-CD47 antibody-mediated elimination of tumor cells? In our own studies, we have shown that the mechanism of antibody action is mediated by macrophages and is due to enabling of phagocytosis by blockade of the CD47-SIRPα that is independent from Fc-receptor (FcR)-mediated antibody killing. Several lines of evidence support this finding. First, this therapeutic effect was primarily mediated by phagocytes, as depletion of macrophages in AML-engrafted immunocompromised mice with liposomal clondronate abrogated the anti-CD47 therapeutic response. Second, two distinct anti-CD47 antibodies that block the CD47-SIRPα interaction enabled phagocytosis whereas a non-blocking anti-CD47 antibody did not, even though all three antibodies bound to target cells similarly. Third, an anti-SIRPα antibody also enabled phagocytosis of AML LSC. Fourth, an isotype-matched anti-CD45 antibody did not enable phagocytosis of LSC even though the antibody bound to the cells. Lastly, an F(ab)’2 fragment of the anti-CD47 antibody enabled phagocytosis of cancer cells even though the Fc receptor was absent (Chao et al., unpublished observations). Thus, these findings support a novel FcR-independent mechanism of antibody action for an anti-CD47 antibody.
Although we have demonstrated that an anti-CD47 antibody eliminates malignant cells by enabling of phagocytosis through inhibition of SIRPα, other mechanisms of anti-CD47 antibody killing have been proposed. Most notably, several groups have shown that anti-CD47 antibodies can eliminate tumor cells by direct apoptosis68,71 that is caspase-independent66. These reports raise the hypothesis that anti-CD47 antibodies induce apoptosis in tumor target cells that are then secondarily phagocytosed by macrophages and other phagocytes and that this process could be independent from the CD47-SIRPα interaction. However, these studies observed apoptosis only when anti-CD47 antibodies were immobilized or cross-linked, whereas apoptosis was not observed when antibodies were incubated in suspension. In our studies, anti-CD47 antibodies enabled phagocytosis of leukemia cells when incubated in suspension with no observed apoptosis59. The functional outcomes in the manner in which CD47 is engaged (cross-linked or in suspension) may be a result of separate signaling mechanisms. In addition to SIRPα, CD47 binds, thrombospondin-1, which forms a molecular bridge between phagocytes and apoptotic cells through its interaction with αvβ3/CD36 72. It is possible that engagement of CD47 in some contexts induces apoptosis (via binding to thrombospondin-173, after irradiation74, or with cross-linking antibodies66), while in other contexts (anti-CD47 antibodies in suspension) blockade of CD47 enables phagocytosis by disrupting CD47-SIRPα. However, it is necessary to further elucidate the biochemical and functional downstream signaling consequences of cross-linked or soluble anti-CD47 antibody-mediated action.
CD47 is a widely expressed protein on many cells of the hematopoietic system as well as other tissues34. Although much evidence supports the pre-clinical efficacy of anti-CD47 antibody-mediated elimination of tumor cells, given the widespread expression of CD47 in the hematopoietic system and other normal tissues, toxicity could be a barrier to translating the pre-clinical efficacy of an anti-CD47 antibody into a clinical therapy. To address this, the effect of an anti-CD47 antibody on normal cells was investigated on both human and mouse normal cell counterparts59. Relatively minimal toxicity was observed. First, normal human CD34+ cells were not phagocytosed in vitro with an anti-CD47 antibody despite phagocytosis of human leukemia cells. Of note, CD34+ cells isolated from cord blood were shown to be resistant to anti-CD47 antibody-induced cell death75. Second, when an anti-mouse CD47 antibody capable of enabling phagocytosis was administered to wild-type mice in vivo, minimal toxicity was observed with no reduction in the HSC pool.
If CD47 is expressed on multiple cell types in the hematopoietic system and in several organ tissues, why is significant toxicity with antibody-targeting of CD47 not observed? The therapeutic selectivity may be due to the presence of a yet unknown pro-phagocytic signal(s) on leukemia cells that is not present on normal cell counterparts, as discussed above. Thus, anti-CD47 antibody therapy may operate in a therapeutic window that is reliant on the presence or absence of pro-phagocytic signals in combination with the anti-phagocytic signal, CD47 (Figure 1e-g).
That macrophages play an important role in tumor immunosurveillance and clearance in myeloid leukemias is underappreciated. However, the notion that macrophages are major mediators of an anti-CD47 antibody therapeutic effect has important implications in translation of such a therapy to human leukemia patients. Standard therapy for human AML involves cytostatic drugs such as anthracycline-based combination chemotherapy which leads to long-term survival rates ranging from 10-75%76, arguing for a need for improved therapy. Chemotherapy has been shown to induce an inflammatory response that attracts infiltrating macrophages into tumors, which in multiple tumors is associated with a better clinical prognosis77-79. Thus, an anti-CD47 antibody therapy administered in the post-chemotherapy setting may result in increased efficacy by utilizing the infiltrating macrophages as immune effector cells at the site of disease. However, if chemotherapy led to long-term expression of pro-phagocytic signals in normal hematopoietic stem and progenitor cells, the anti-CD47 therapy could have significant added hematopoietic toxicity.
In addition to modulation by chemotherapy, increasing macrophage cell numbers may also augment the efficacy of an anti-CD47 antibody therapy by expanding the available pool of effector phagocytic cells. Such macrophage manipulation could be achieved by increasing endogenous numbers of macrophages through administration of cytokines including macrophage-colony stimulating factor (M-CSF) or granulocyte-macrophage-colony stimulating factor (GM-CSF). The combination of such cytokines with tumor-specific antibodies has been shown to be therapeutically effective in tumor models80,81. In addition, GM-CSF is clinically approved for use in AML patients and is generally well-tolerated (reviewed in 82). Therefore, combination therapy with GM-CSF and anti-CD47 antibody could be a viable therapy that could be translated clinically. In addition to increasing endogenous macrophage production, exogenous macrophages could be increased through ex vivo transplantation of macrophage progenitors including chronic myeloid progenitor (CMP) and granulocyte-macrophage progenitor (GMP) cells, which could be followed by anti-CD47 antibody administration to form an effective therapeutic strategy. We have previously shown that transplantation of either syngeneic83 or fully allogeneic84 mouse CMP/GMP into mice challenged with bacterial or fungal pathogens can increase myeloid effector cell numbers and prevent fulminant infection and death. Currently, we are investigating these therapeutic combinations in pre-clinical mouse models.
Evidence is emerging that for human myeloid leukemias, macrophage phagocytosis is a critical mediator of tumor immunosurveilance. These cancers can escape such immunosurveillance by up-regulating CD47, a potent “don’t eat me” signal. This appears to be a mechanism co-opted from normal HSCs, which also up-regulate CD47 as protection against phagocytosis in response to inflammatory or mobilizing agents. We are currently exploring the ability of blocking anti-CD47 antibody in combination with agents that increase macrophage activity to treat human myeloid leukemias in pre-clinical models.
The authors would like to thank Maureen Howard for critical evaluation of the manuscript. S.J. is supported by the Stanford Medical Scientist Training Program. M.P.C. is supported by a Medical Student Training Fellowship from the Howard Hughes Medical Institute. R.M. is supported by a grant from the American Association for Cancer Research, and holds a Career Award for Medical Scientists from the Burroughs Wellcome Fund. I.L.W. is the Virginia and Daniel Ludwig Professor of Clinical Cancer at Stanford University. I.L.W. is an inventor of U.S. Patent Application 11/528,890 entitled “Methods for Diagnosing and Evaluating Treatment of Blood Disorders”. I.L.W., S.J., M.P.C., and R.M. have filed U.S. Patent Application Serial No. 12/321,215 entitled “Methods For Manipulating Phagocytosis Mediated by CD47”. This research is supported by National Institutes of Health Grants 5R01CA086017-08, 5R01HL058770-08, 5P01DK053074-08 awarded to I.L.W., the de Villier award of the Leukemia Society, the Ludwig Institute, and a gift from the Smith Family Fund.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.