| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print June 3, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,,1
* INSERM U457 and
¶ Paediatric Endocrinology Department, Robert Debré Teaching Hospital, Paris, France;
EMI0363 Necker University, Paris, France;
Mount Sinai School of Medicine, New York, NY;
Paediatric Endocrinology Department, Necker-Enfants-Malades Teaching Hospital, Paris, France; and
|| Albert Einstein College of Medicine, Bronx, NY
1Correspondence: Service d’Endocrinologie et Diabétologie Pédiatriques, Hôpital Necker Enfants-Malades, 149 rue de Sèvres, 75015 Paris, France. E-mail: michel.polak{at}nck.ap-hop-paris.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: pancreas • development • remodeling • pregnancy • osteopetrotic mice
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Endocrine cell mass increases throughout life. In the rat fetuses, islet cell mass increases dramatically as a result of insulin cell replication and of recruitment and maturation of undifferentiated insulin-cell precursors within the pancreatic ducts. After birth, the growth rate of all islet cells starts to decline within 3–4 days and continues to decrease thereafter until adulthood, when the rate is low [11 12 13 ]. Important pancreatic islet remodeling has been shown in neonatal rats [14 15 16 ] and pregnant females. Maternal pancreatic adaptation during pregnancy involves profound morphological and functional changes [17 18 ].
The aim of this work was to look for the presence of F4/80-positive macrophages in wild-type pancreas throughout life and if such cells were found to perform studies looking for their involvement in pancreas development, growth, and remodeling during fetal and postnatal life. We use a murine model homozygous for a spontaneous null mutation in the gene encoding colony-stimulating factor 1 [CSF1; Csf1op/Csf1op (op/op)]: This mutation results in production of a truncated, nonfunctional protein [19 20 ]. CSF1, known to be synthesized by various cell types (epithelial, mesenchymal, endothelial) [21 ], signals only through a cell-surface tyrosine-kinase receptor present on all cells of mononuclear phagocyte lineage (CSF1R) [22 23 24 ]. Phenotypically op/op mice are osteopetrotic, toothless, and have a reduced macrophage number in most of their tissues [25 ]. Their pancreas has not been investigated. op/op mice are a useful tool for studying the macrophage role in organ development [26 27 28 ] and function [29 ].
We found macrophages to be present in the developing pancreas in mice. We showed that macrophage-deficient op/op mice lacked CSF1-dependent macrophages in the pancreas. Hence, we aimed this work to evaluate the endocrine-cell mass, study the insulin cell number, size, and organization within islets, and understand the mechanisms of pancreas development/growth and remodeling throughout life in macrophage-deficient op/op mice.
Relying on the data we obtained while comparing op/+ controls and op/op mice, we do propose that CSF1-dependent macrophages are essential for full insulin cell mass development, postnatal islet morphogenesis, and pancreatic cell proliferation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We studied the macrophage role in pancreas development using osteopetrotic Csf1op/Csf1op, thereafter op/op [19 20 30 ], and littermate control (op/+) mice obtained by mating op/+ females and op/op males and maintained in a barrier facility at the Albert Einstein College of Medicine (Bronx, NY), as described by Pollard and Hennighausen [28 ]. The op/op mice were fed powdered chow ad libitum and infant milk formula (Enfamil, Mead Johnson and Co., Evansville, IN). We focused our study on time-points when major morphological and functional changes occur in the pancreas namely on day 18.5 post-coitum (E18.5); weanings, adults and pregnant adults, at 18.5 dp.
CSF1 crosses the placenta and is required for the perinatal development of most macrophages in the liver, spleen, kidney, and lung. Circulating maternal CSF1 can compensate for the absence of CSF1 production by op/op fetuses and ensure normal macrophage development in the liver but not in the spleen, kidney, or lung [31 32 ]. As no information is available about the putative role of maternal CSF1 on macrophage development in the embryonic or fetal pancreas, we compared op/+ embryos from op/+ mothers to op/op embryos from op/op mothers. op/op mothers are unable to feed their pups, as they do not develop normal mammary glands. Consequently, postnatal data in our study were collected from op/op mice obtained by mating op/op males and op/+ females and suckled until the normal weaning time.
To allow pancreatic cell proliferation study, the animals were killed 2 h after intraperitoneal administration of 5-bromo-2-deoxyuridine (BrdU; 100 µg/mg); the pancreas and kidneys were then dissected, weighed, fixed in formalin, and embedded in paraffin.
Fetal genotype and sex at E18.5 were determined by polymerase chain reaction on tail DNA. After birth, animals with the op/op genotype were identified based on the absence of incisors.
Sectioning.
Each paraffin-embedded pancreas was cut into 4–5 µm coronal sections in its entirety. Because of pancreatic endocrine and proliferation heterogeneity [33
], the study was done on five sections (for postnatal pancreas) to 10 sections (for fetal pancreas), sampled at regular intervals throughout each pancreas.
Histological analysis: Haematoxylin-eosin staining was performed on all the studied organs to evaluate tissue histology. No abnormalities were seen in the op/op mutants or op/+ controls.
Immune staining of paraffin-embedded sections.
Macrophage identification: To look for macrophages, we used F4/80 antibody against a murine macrophage-restricted, cell-surface glycoprotein marker [34
].
CSF1R expression was studied on pancreases from E18.5 fetal Swiss mice (fixed in PLPG). We used two different antibodies and three different incubation conditions [1 h at 37°C, 5 h at room temperature, and overnight at 4°C] to improve the sensibility of CSF1R protein detection.
Endocrine cells: To evaluate the expression of endocrine markers (insulin and glucagon), we used diaminobenzidine as a chromogen to allow for quantitative morphometric analysis as described previously [33 ], and on the adjacent section, we performed a double staining with fluorescein isothiocyanate (FITC) and Texas Red fluorescent chromogens to identify insulin- and glucagon-expressing cells and to look for hormone coexpression.
Cell proliferation: Incorporation of BrdU into proliferating nuclei was detected by fluorescent immunohistochemistry, allowing a double staining with insulin to identify proliferating endocrine cells (see Fig. 2C 2D 2E 2F ).
|
|
Controls.
Staining specificity was controlled on sections of normal adult mouse pancreas for the endocrine markers, mouse liver for macrophage marker expression, and mouse lymph node for BrdU incorporation into the nuclei of proliferating cells (see Fig. 2H
).
TdT-mediated Bio 11 deoxy-uridine 5'-triphosphate nick end-labeling coupled to digoxigenin (TUNEL) analysis: Apoptosis was studied by TUNEL using the ApopTag kit (Qbiogene, Carlsbad, CA). For prenatal studies and when needed to achieve islet detection, double immunostaining with insulin or glucagon was done to look for colocalization of apoptotic nuclei and cytoplasm endocrine markers.
We used human colon follicles and mouse lymph nodes as control tissues for evaluating apoptosis. The specificity of apoptotic nuclei staining was assessed based on morphology. Cross-reaction with proliferating nuclei has been reported [35 ] but did not occur in our study, as shown by the finding that the number of apoptotic nuclei was usually 100 times smaller than the number of proliferating nuclei in the adjacent section.
Quantitative data
Quantitative morphometric analysis
The microscopic fields acquired from insulin, immune-stained, pancreatic sections were measured using a Leica DMRB microscope equipped with a color video camera, connected to a Quantimet 500 IW computer (screen magnification, x10), as described previously [33
36
].
The number of islets (defined as insulin-positive aggregates at least 25 µm in diameter) was scored and used to calculate the islet numerical density (number of islets per square millimeter of pancreatic tissue). Islets ranging from 25 to 100 µm in diameter were defined as small, those ranging from 101–150 µm as medium-sized, and those exceeding 150 µm (151–600 µm) as large. Postnatal pancreases were used for this study, as fetal pancreas contained aggregated endocrine cells not grouped in recognizable islets (see Fig. 2C and 2D ). The percentages of small, medium-sized, and large islets relative to the total number of islets were calculated.
Endocrine mass was calculated as previously explained [36 37 ] and normalized for total body weight. Approximate insulin cell size (in µm2) was determined by dividing insulin surface area by insulin cell number.
Cell quantification
Insulin-positive cells were counted under a fluorescent microscope. Counts per pancreas were as follows: 229–881 cells at E18.5, 1745–2996 on postnatal day 21, 1573–7595 in nonpregnant adults, and 3135–6171 in pregnant adults. Postnatal islets identified based on insulin-positive staining were counted under the microscope. The number of islets per pancreas was 95–197 on postnatal day 21, 112–182 in nonpregnant adults, and 109–179 in pregnant adults. The number of islets budding from or close to the sectioned ducts was determined (see Fig. 2A
and 2B
). BrdU-positive proliferating cells were counted, the number per pancreas being 1240–3160 cells at E18.5, 77–4517 on postnatal day 21, 113–643 in adults, and 39–2532 in pregnant adults. BrdU-positive duct cells (see Fig. 2E
and 2G
) and insulin-proliferating cells were counted.
The total number of apoptotic nuclei (10–68 cells per pancreas) and their percentage relative to the total number of nuclei were determined in connective, exocrine, and endocrine tissues.
All islet and cell numbers (insulin-positive, proliferating, and apoptotic) were normalized for pancreas surface area, and results were expressed per pancreas surface unit.
Values of P smaller than 0.05 were considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
CSF1R staining on PLPG-fixed pancreases from E18.5 Swiss mice showed that isolated dendritic-like, CSF1R-positive cells were present only in the connective tissue (Fig. 1D and inset). No CSF1R expression was documented in the endocrine cells (aggregated at this stage, Fig. 2C ), the acinar cells (Fig. 1D and inset), or the duct epithelial cells (data not shown). The morphology and location of these cells (Fig. 1D inset) were similar to those of F4/80-positive macrophages (Fig. 1A and 1B and inset).
CSF1-dependent macrophages positive for F4/80 [34 ] and/or CSF1R [22 23 24 ] markers were detected throughout the mesenchyme/connective tissue of the pancreas from E14.5 onward in Swiss and op/+ mice. In op/op mice pancreas, macrophages were absent or severely depleted at all the studied stages.
Insulin cell mass is altered in the absence of macrophages
Weight analysis
The op/op mice showed growth retardation throughout their lifetime, as compared with op/+ controls except for pregnants (Table 2
). Normalized op/op pancreas and kidney weights as compared with controls were not statistically significant, excepted at weaning time when relative pancreas weight was significantly increased, and relative kidney weight was significantly decreased (Table 2) .
|
Endocrine cell mass
Pancreatic surface area was closely correlated to total pancreatic weight [coefficient of correlation between mean pancreatic surface area (µm2) and pancreas weight throughout life, 0.99 for op/+ mice and 0.98 for op/op mice], confirming that the sampling method used in our study is valid for endocrine cell mass evaluation [33
36
37
].
In op/op mice, insulin cell mass was decreased in the fetuses and subsequently continued to expand at a reduced rate (Table 3 ) as compared with the controls. Normalized op/op insulin cell mass was significantly decreased in the fetuses and adults with a similar trend, although not statistically significant at weaning (Table 3) .
|
Analysis of insulin tissue
Insulin cell number and size.
Insulin tissue develops by division (which increases cell number), hypertrophy (which increases cell size) of differentiated insulin cells, and/or by neogenesis (differentiation of putative epithelial duct precursors) [12
13
14
38
]. We first analyzed insulin cell number and size to determine whether cell scarcity and/or hypotrophy explained the insulin cell mass decrease in the op/op mice.
In op/+ mice, the insulin cell number per mm2 pancreas surface area was halved postnatally compared with fetal values (Fig. 2C) , as a consequence of the dramatic exocrine tissue expansion compared with the endocrine tissue expansion. No significant differences were found in insulin cell numbers per mm2 between controls and age-matched op/op mice at any of the time-points studied (data not shown).
Insulin cell size (in µm2) was decreased in op/op mice as compared with controls at E18.5 (74±21 vs. 164±4, P=0.019) and on postnatal day 21 (107±31 vs. 154±38, P=0.037); in contrast, no significant difference was seen for adult (176±39 vs. 303±88, P=0.07) and pregnant (168±70 vs. 203±31, P=0.5) mice. These findings suggested that insulin cell hypotrophy may have contributed to the decrease in fetal insulin cell mass but that compensation by proliferation and/or neogenesis occurred in op/op mice at weaning.
Postnatal islet morphogenesis
Islets were present in the op/op mice and the controls (Fig. 2A 2B
2E
2F)
. Insulin was expressed in the islet core and glucagon in the islet periphery in both groups (data not shown). Coexpression of insulin and glucagon was not seen at any of the time-points (data not shown).
Islet number, size, and location.
In postnatal op/+ mice, islet density (number per mm2 pancreatic surface area) was significantly greater on postnatal day 21 than in pregnant adults (P=0.049). As compared with controls, the islet number per mm2 in the macrophage-deficient pancreases was significantly increased on postnatal day 21 (2.55±0.5 vs. 1.9±0.27, P=0.01) and not different in adults (1.65±0.3 vs. 1.16±0.17, P=0.08) and in pregnant (1.35±0.3 vs. 1.22±0.08, P=0.5) mice. As the insulin cell number per mm2 was similar, that suggested a different insulin cell organization within islets in op/op mice at weaning.
To confirm this, we then counted small, medium-sized, and large islets. Small islets were significantly more numerous on postnatal day 21 in op/op mice (Table 4 ). Large islets were significantly less numerous at all time-points in op/op mice. These results suggested islet growth impairment at all postnatal time-points in the op/op mice. To determine whether the increased number of small islets in op/op mice at weaning reflected neogenesis, we examined islet location relative to the ducts.
|
On postnatal day 21 in op/op mice, increased neogenesis from ducts, suggested by the increase of small islets and islets budding from ducts, could counterbalance the insulin cell hypotrophy and impaired large islet formation. This may explain why the insulin cell mass at weaning in op/op mice is not statistically different from controls.
To elucidate the mechanisms underlying our findings, we investigated proliferation and apoptosis, which together with neogenesis, are the mechanisms of insulin cell mass remodeling.
Pancreatic proliferation and apoptosis
Proliferation
Total pancreatic proliferation.
In controls, pancreatic cell proliferation decreased dramatically after birth, as compared with the fetal time-points, and increased during pregnancy (Fig. 3A
).
|
Duct cell proliferation.
As exocrine and endocrine precursors are considered to originate in ducts [40
41
], we used the percentage of proliferating cells located in ducts as an indicator of the potential precursor pool. We performed this analysis only in postnatal pancreases, in which the ducts were numerous and identified based on their particular histological features (Fig. 2A
2B
2E
2G)
. The percentage of proliferating epithelial duct cells was increased only at postnatal day 21 in pancreases from macrophage-deficient mice as compared with controls (Fig. 3B)
. This suggested that the ductal precursor pool was increased at weaning. This finding was in keeping with the increased islet neogenesis found in our study at this stage.
Insulin cell proliferation.
We examined insulin cell proliferation, which is a mechanism of insulin cell mass expansion [12
38
]. The percentage of proliferating insulin cells was higher in the fetuses (Figs. 2C and 3C)
than after birth. Compared with controls, it was dramatically decreased in op/op fetal pancreases (Figs. 2D
and 3C) , whereas no differences were seen between the two groups at the postnatal time-points (Fig. 3C)
.
Therefore in op/op fetuses, the impaired insulin mass expansion is the result of the major insulin proliferation reduction combined with the insulin cell hypotrophy.
Apoptosis analysis
In controls, the number of apoptotic nuclei per mm2 pancreas was smaller after than before birth (Fig. 3D)
. Apoptotic cells were located chiefly in the connective tissue prenatally and in the acinar tissue postnatally. In macrophage-deficient pancreases, apoptotic cells were more numerous at all time-points than in the controls (Fig. 3D)
and chiefly located in the acinar tissue. Endocrine cell apoptosis was rarely detected, even in macrophage-deficient pancreases with increased apoptotic cell counts.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To analyze the role of CSF1-dependent macrophages in pancreatic development, we studied this process in op/op mice, and the data obtained are summarized in Table 5 .
|
After birth, insulin cell number per body weight decreases with age [14 ]. Neogenesis occurs during suckling, as insulin-cell mass did not change the first 2 weeks of life despite the reported wave of apoptosis [16 ]. In op/op mice at 3 weeks of postnatal development, insulin-cell mass and proliferation have reached values not significantly different from controls, although hypotrophy persists, and the number of large islets is decreased. This probably results from an increased neogenesis, as islet density and islets budding from ducts are significantly increased [46 47 48 ]. Acinar cell mass contributes to 95% of the pancreatic mass [2 ]. A striking pancreatic, mainly acinar, impaired proliferation was demonstrated at weaning time and in pregnant osteopetrotic mice despite normal pancreatic relative weight, suggesting the possibility of a pancreatic cell mass compensation by the increase in acinar cell size and/or differentiation. Neogenesis and acinar differentiation may relate to our finding that proliferating epithelial duct cell number is markedly increased at weaning, suggesting occurrence of an endocrine and exocrine precursor pool expansion [2 ].
Between weaning time and adult stage, insulin-cell proliferation leads to an increased mass of two to four times by islet enlargement [14 49 ]. Insulin-cell proliferation is decreased in adult mice, and insulin-cell renewal is low [13 50 ]. In adult op/op mice, insulin-cell mass is decreased. Large islets are less numerous, leading to a decreased insulin volume. Therefore, after the weaning, pancreatic compensation has failed to maintain a normal insulin-cell mass to adulthood.
Morphological, pancreatic adaptation to pregnancy is achieved by 1.4- to threefold islet mass increase, by insulin and glucagon-cell proliferation and insulin cell hypertrophy, contributing to maternal physiological hyperinsulinemia necessary for placental and fetal growth [51 ]. Insulin cell mass is abnormally low in pregnant op/op mice. The number of large islets is decreased, resulting in a decrease in insulin volume [52 ]. Thus, maternal pancreatic morphological adaptation to pregnancy does not take place in the absence of CSF1-regulated macrophages.
Apoptotic cell density was increased in the fetal pancreatic mesenchyme and in the postnatal acinar tissue. This is consistent with the known constitutive function of macrophages in clearing apoptotic cells [53 ], as recently described in the mouse thymus [54 55 ] and shown during mouse mammary gland development [26 ]. However, endocrine cell apoptosis was not detected in our study at all the studied time-points [38 ]. Therefore, compensation by other phagocytic cells [53 56 ] may have occurred. The clearing process of these phagocytes could be less efficient than that of macrophages, resulting in an increased number of detected apoptotic cells [56 ].
op/op mice present intra-uterine growth retardation that persists throughout life. Malnutrition may affect pancreatic development. In different malnutrition models, impaired fetal insulin cell number and mass were not associated with insulin cell size and proliferation defect [36 ] or were associated with proliferation defect and increased islet apoptosis [57 ]. The outcome in adults was glucose intolerance or gestational diabetes. op/op adults and pregnants do not present hyperglycemia (data not shown). By contrast, in the macrophage-deficient mice model, body weight retardation in malnutrition models is not reversible during gestation [36 37 58 59 ]. Therefore, the findings in op/op mice are clearly at variance with those obtained in murine models of malnutrition and suggest that the data are a result of the absence of macrophages.
In vitro studies demonstrated that the mesenchyme is involved in epithelial development, growth, and differentiation [7 10 ]. By secretion of soluble factors, it represses endocrine differentiation and stimulates exocrine proliferation and differentiation [7 10 ]. In vivo, endocrine differentiation occurs in the presence of the mesenchyme, which raises the question of a complex regulation of the epithelium-mesenchyme interaction in pancreas development [33 ].
CSF1-dependent macrophages are present in the pancreatic mesenchyme/connective tissue. Indirect evidence for neogenesis [11 15 39 46 47 60 ] was observed at all the studied time-points in op/op mice. Macrophages could therefore be involved in neogenesis repression, according to the mesenchymal role documented in vitro [7 ]. Alternatively, neogenesis may occur as a mechanism that compensates for impaired insulin cell mass development, as seen in insulin cell regeneration models [14 61 62 ]. Profound proliferation impairment was found in op/op mice at weaning and at late pregnancy, suggesting that macrophages and/or the factors they secrete may be involved in morphological tissue remodeling and adaptation occurring at these stages [29 44 63 ]. In keeping with this possibility, in vitro studies suggest a role for mesenchymal factors in exocrine proliferation and differentiation [7 10 ].
In conclusion, we present a detailed analysis of pancreatic development in a macrophage-deficient mouse model, showing for the first time that resident CSF1-dependent macrophages are implicated in insulin cell mass expansion by exerting effects on large islet morphogenesis, pancreatic cell proliferation, and insulin cell neogenesis. The macrophage role in pancreatic development, growth, and weaning and pregnancy-related processes should now be confirmed by in vitro experiments.
|
ACKNOWLEDGEMENTS |
|---|
Received February 18, 2004; revised March 31, 2004; accepted April 15, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) system studied using the colony-stimulating factor-1-deficient op/op mouse Physiol. Rev. 76,927-947This article has been cited by other articles:
|
|
 |
|
  J. S. Tessem, J. N. Jensen, H. Pelli, X.-M. Dai, X.-H. Zong, E. R. Stanley, J. Jensen, and J. DeGregori Critical Roles for Macrophages in Islet Angiogenesis and Maintenance During Pancreatic Degeneration Diabetes, June 1, 2008; 57(6): 1605 - 1617. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Calderon, A. Suri, M. J. Miller, and E. R. Unanue Dendritic cells in islets of Langerhans constitutively present {beta} cell-derived peptides bound to their class II MHC molecules PNAS, April 22, 2008; 105(16): 6121 - 6126. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Shaykhiev and R. Bals Interactions between epithelial cells and leukocytes in immunity and tissue homeostasis J. Leukoc. Biol., July 1, 2007; 82(1): 1 - 15. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. B. Geutskens, T. Otonkoski, M-A. Pulkkinen, H. A. Drexhage, and P. J. M. Leenen Macrophages in the murine pancreas and their involvement in fetal endocrine development in vitro J. Leukoc. Biol., October 1, 2005; 78(4): 845 - 852. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
