The Soluble Type 2 Insulin-like Growth Factor (IGF-II) Receptor Reduces Organ Size by IGF-II-mediated and IGF-II-independent Mechanisms*

The soluble type 2 insulin-like growth factor (IGF) receptor or IGF-II/mannose 6-phosphate receptor (sIGF2R) is produced in vivo by proteolytic deletion of the transmembrane and intracellular domains of the cellular form of the receptor (IGF2R). There is evidence that sIGF2R is a negative regulator of growth. We have shown that transgenic mice expressing anIgf2r cDNA with a deleted transmembrane domain sequence (sΔIgf2r) show reduced local organ size. In the present study, we investigate whether sΔIGF2R can slow the growth induced by an excess of IGF-II and whether the biological activity of sΔIGF2R is due solely to its interactions with IGF-II. To this end, we crossed sΔIgf2r transgenics by mice overexpressing IGF-II (Blast line) or by mice carrying a disrupted paternal (active) allele of the Igf2 gene (Igf2 m+/p− ). Analysis of the phenotypes revealed that the soluble IGF2R affects the size of some organs (colon and cecum) exclusively by reducing the biological activity of IGF-II, whereas in other organs (stomach and skin) the biological activity of the receptor is at least in part independent of IGF-II and must involve an interaction with other factor(s).

receptor or IGF-II/mannose 6-phosphate receptor (sIGF2R) is produced in vivo by proteolytic deletion of the transmembrane and intracellular domains of the cellular form of the receptor (IGF2R). There is evidence that sIGF2R is a negative regulator of growth. We have shown that transgenic mice expressing an Igf2r cDNA with a deleted transmembrane domain sequence (s⌬Igf2r) show reduced local organ size. In the present study, we investigate whether s⌬IGF2R can slow the growth induced by an excess of IGF-II and whether the biological activity of s⌬IGF2R is due solely to its interactions with IGF-II. To this end, we crossed s⌬Igf2r transgenics by mice overexpressing IGF-II (Blast line) or by mice carrying a disrupted paternal (active) allele of the Igf2 gene (Igf2 m؉/p؊ ). Analysis of the phenotypes revealed that the soluble IGF2R affects the size of some organs (colon and cecum) exclusively by reducing the biological activity of IGF-II, whereas in other organs (stomach and skin) the biological activity of the receptor is at least in part independent of IGF-II and must involve an interaction with other factor(s).
The growth and survival factor insulin-like growth factor II (IGF-II) 1 binds to at least three different receptors: the type 1 and 2 IGF receptors and IGF2R (also known as the mannose 6-phosphate/IGF-II receptor) and the insulin receptor. Type 1 IGF receptor and the insulin receptor are members of the tyrosine kinase receptors family and mediate most of the biological effects of IGF-II (1,2).
Genetic evidence suggests that the IGF2R gene encodes a negative regulator of growth. Many human tumors show loss of heterozygosity at the IGF2R locus frequently accompanied by mutations in the remaining allele (3)(4)(5). Furthermore, mice in which the Igf2r gene is disrupted are born 25-35% bigger than controls (6 -8). IGF2R is a multifunctional protein that participates in the activation of TGF-␤1, regulates lysosomal enzymes trafficking, and binds a number of ligands including proliferin, herpes simplex virus glycoprotein D, thyroglobulin, and retinoic acid (9 -14).
A soluble form of IGF2R (sIGF2R) is produced by proteolytic cleavage of the transmembrane and intracellular domains of the membrane form of the receptor and is present in the serum, amniotic fluid, and urine of rodents and humans. sIGF2R binds IGF-II with high affinity in vivo and can bind mannose 6-phosphate in vitro, suggesting that it shares at least some of its ligand specificity with the membrane IGF2R (15)(16)(17)(18)(19). There is evidence that sIGF2R is a biologically active molecule. First, sIGF2R can inhibit DNA synthesis induced by IGF-II and epidermal growth factor in cultured rat hepatocytes (20). Second, we have obtained transgenic mice expressing a soluble IGF2R by deletion of the transmembrane domain sequence (s⌬Igf2r) and fused to the regulatory sequence of the keratin 10 promoter to target expression to the alimentary canal, skin, and uterus (K10s⌬Igf2r transgene). Two lines of K10s⌬Igf2r transgenic mice (Kipps and Krishna) showed a 9 -20% reduction of wet weight, dry weight, and water content in the alimentary canal. The effects of s⌬IGF2R expression were mainly local, because the organs negative for transgene expression were only marginally affected (21).
The interpretation of the biological activity of sIGF2R is complicated by the heterogeneity of its ligands. To understand to what extent the biological activity of s⌬IGF2R is due to interaction with IGF-II, we crossed K10s⌬Igf2r transgenics by the following genetically modified mice: 1) transgenics expressing a K10Igf2 minigene and showing local organomegaly (Blast line; Ref. 22). If s⌬IGF2R acts by reducing the activity of IGF-II, double transgenics Blast and K10s⌬Igf2r should show an attenuation of organomegaly compared with Blast. 2) mice in which the paternal (active) allele of the Igf2 gene is disrupted (Igf2 ϩ/pϪ ). Igf2 mϩ/pϪ mice show a growth deficiency phenotype and are fertile (23). If s⌬IGF2R acts exclusively by interacting with IGF-II, organ size should not be affected in K10s⌬Igf2r transgenics that are also Igf2 mϩ/pϪ compared with Igf2 mϩ/pϪ mice. The results of the present work provide insights into the mechanism of organ size reduction by the soluble IGF2R.

EXPERIMENTAL PROCEDURES
Transgenic Mice-K10s⌬Igf2r (lines Kipps and Krishna) transgenic mice express a mutant mouse Igf2r cDNA in which the sequence encoding the transmembrane domain has been deleted to encode a soluble polypeptide (s⌬IGF2R). The mutant cDNA is under the transcriptional control of the keratin 10 promoter (K10). Kipps is the line expressing the K10s⌬Igf2r transgene at highest levels, and Krishna is the second best expressing line; their phenotype has been described in detail (21). K10Igf2 (Blast) and Igf2 mϩ/pϪ mice have been described elsewhere (22,23). All mice used in this study were heterozygotes (indicated as K10s⌬Igf2r/ϩ, K10Igf2/ϩ and Igf2 mϩ/pϪ ) in a mixed genetic background as the integrated transgenes were bred from a F 1 (C57Bl/6 ϫ CBA) onto a 129J/Sv genetic background. In both crosses the K10s⌬Igf2r transgene was transmitted maternally, because Blast fe-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Analysis of Organ Parameters-Analysis of organ weights was performed in 90-day-old mice produced by crossing heterozygotes K10s⌬Igf2r/ϩ (Kipps or Krishna line) by K10Igf2/ϩ or Igf2 mϩ/pϪ mice. Live weight was recorded, and then anesthetized mice were bled by decapitation and dissected, and the total wet weight of organs was measured after removal of fat and mesenteries. The contents of the alimentary canal were removed by gentle scraping in phosphate-buffered saline followed by blotting in tissue paper. Organs were always dissected in the same order to normalize wet weight loss due to evaporation. Each organ was cut in two or three parts, and the parts were weighed. This made it possible to calculate the total organ dry weight, water content, DNA, and detergent-soluble protein content after assaying different parts.
For dry weight measurement, tissue fragments were dried at 65°C for 6 days. Detergent-soluble protein was measured by the Coomassie Blue method (Bradford reagent, Sigma) after tissue homogenization in 20 mM Tris, pH 7.5, 10 mM EDTA, 0.1% Tween 20, 0.15 M NaCl. DNA content was measured in the same extract usingt the Hoechst 33258 fluorochrome, following addition of NaCl to 2 M final concentration and brief sonication (25). For all measured parameters, relative values (e.g. per mg wet weight) were first obtained by dividing the value obtained for a given tissue part by the weight of that tissue part. This relative value was then multiplied by the total organ wet weight to obtain the total value (e.g. for the whole organ). Total water content was calculated by subtracting the total dry weight from the total wet weight. Total water content was divided by the total dry weight to obtain the water content/unit dry weight.
All comparisons were by paired t test of pairs of mice matched for litter, age, and sex. The number of paired comparisons used in each t test was in some cases lower than the total number of animals displayed in each column of genotypes (Tables I-IV). The reason was that in comparing two given groups, only mice that could be matched were used, and this was frequently a subset of one of the two groups. Thus the description of the phenotype for each genotype given in the tables does not always indicate the mean value of the phenotype character that was involved in paired t tests.
RNA Analysis-Igf2 RNA was analyzed by reverse transcription-PCR. 1 g of total RNA was reversed transcribed by using the Reverse Transcription System according to the manufacturer's instructions (Promega). cDNA was amplified by PCR using the primers 5Ј-GAAGT-CGATGTTGGTGCTTCTCATCTC-3Ј (forward) and 5Ј-GACAAGCCTG-GCGCCGAAGATGAAGT-3Ј (reverse). These primers recognize sequences in exon 4 of the Igf2 gene and produce a 139-bp product (Ref. 22 and references therein). To control for errors in pipetting and RNA quantitation, primers specific for the mouse G3PDH cDNA were included in the same PCR tube (CLONTECH). The conditions for amplification were 94°C for 3 min, 53°C for 2 min, and 72°C for 2 min for 35 cycles. Products were separated in a 1.6% TBE-agarose gel and stained with ethidium bromide.

RESULTS
Crosses K10s⌬Igf2r/ϩ ϫ K10Igf2/ϩ-To find out whether the soluble receptor could reduce growth stimulated by excess local IGF-II, two transgenes were crossed into the same mice. The transcription of both transgenes was driven by the same keratin 10 promoter, and their patterns of organ expression overlapped. The matings to produce these double transgenics also generated mice with one or the other of the transgenes. The phenotype of the single K10Igf2/ϩ transgenic is first described to provide the base line for comparison with the double transgenic.
Heterozygous transgenic mice expressing the Igf2 gene under the transcriptional control of the keratin 10 promoter (Blast line, K10Igf2/ϩ) showed local organ overgrowth as described previously (22). Organ wet weight was significantly increased in the alimentary canal, skin, and uterus compared with controls, whereas wet weight was normal in organs negative for transgene expression (column A versus column B in Tables I and II). Total dry weight followed a very similar pattern in the organs examined (column A versus column B, Tables I and II). An increase of water content relative to dry matter content (edema) has been shown in mice with elevated levels of IGF-II (8). The water content relative to dry weight was unchanged in Blast in all organs examined (columns A versus column B, Tables I and II). DNA and detergent-soluble protein contents did not change significantly in any of the organs examined, with the exception of the colon as reported (22). In this organ the total DNA content was increased by a similar extent as wet and dry weights (Table I, column A versus column B).
The phenotype of K10s⌬Igf2r/ϩ mice was a reduced wet weight of the alimentary canal (21). The aim of the cross K10s⌬Igf2r/ϩ ϫ Blast was to compare the extent of organomegaly in double transgenics with wild type and Blast/ϩ. For these two reasons, the phenotype of K10s⌬Igf2r/ϩ mice will be described in the context of the cross K10s⌬Igf2r/ϩ ϫ Igf2 mϩ/pϪ (see the next section; column C in Tables III and IV).
In the double transgenics, the extra soluble receptor transgene (K10s⌬Igf2r, either Kipps or Krishna) was in the same mouse as the transgene that expressed excess IGF-II (K10Igf2, Blast line). A reduction of organomegaly was observed in most of the organs coexpressing the two transgenes. Organ wet weight was significantly reduced in double transgenics compared with Blast in the alimentary canal (column C versus column B, Table I). This reduction ranged from more than 100% in the stomach to about 50% in the cecum and colon and was more pronounced in the Blast/Kipps double transgenics (column C, Table I; Fig. 1). This result is consistent with the relative levels of the K10s⌬Igf2r transgene expression in the two lines (21). In the skin, a significant reduction of organomegaly (ϳ50%) was observed only in Blast/Kipps double transgenics (column C versus column B, Table I). No effect of s⌬IGF2R expression was observed on the wet weight of the uterus (column C versus column B, Table I).
Coexpression of the two transgenes produced a decrease in organ dry weight comparable with the decrease in wet weight and did not alter DNA content or detergent-soluble protein content (column B versus column C, Table I). The decrease in dry weight was more marked in Blast/Kipps than in Blast/ Krishna double transgenics (column C, Table I). DNA content was significantly reduced in the small intestine and colon in Blast/Krishna mice compared with Blast (column C versus B, Table I). No major change in the organ parameters measured was observed in the organs, which did not express the transgene (column C versus column B, Table II).
Crosses K10s⌬Igf2r/ϩ ϫ Igf2 mϩ/pϪ -Mice carrying a disrupted Igf2 paternal allele (Igf2 mϩ/pϪ ) were 30% smaller than wild type at 90 days in a C57Bl/6/CBA/129J/Sv mixed genetic background and were fertile as reported (Ref. 23; column B versus column A, Table III). The levels of IGF-II transcript and peptide in Igf2 mϩ/pϪ mice were below the detection limit in three independent assays. First, no IGF-II transcript was detectable in the stomach at 10 days by reverse transcription-PCR (results not shown). Second, serum obtained from Igf2 mϩ/pϪ mice was included in a IGF-II radioimmunoassay as a negative control during the characterization of the K10s⌬Igf2r/ϩ transgenics (21). IGF-II peptide levels were below the sensitivity of the assay. Third, in an independent study in our laboratory, the stomach and small intestine of Igf2 mϩ/pϪ mice were shown to contain no detectable Igf2 transcript when analyzed by a specific RNA in situ hybridization assay. 2 Analysis of individual organs revealed that the wet weight of most organs was proportionate to the whole body weight (values of r close to 1, column B in Tables III and IV). The stomach and the skin were exceptions, because their wet weights were disproportionately large in relation to the whole body weight (values of r larger than 1, column B in Table III). Total organ dry weights followed a pattern similar to wet weights in the 2 B. Hassan, unpublished observations.  organs examined, whereas no change was observed in water content relative to dry weight, DNA or detergent-soluble protein contents (column B versus column A, Tables III and IV). K10s⌬Igf2r/ϩ mice showed a reduction of the wet weight of the stomach, cecum, and colon ranging from 14 to 30% compared with wild type. The skin was not affected, and the liver was the only organ not expressing the transgene that showed a reduction in size (column C in Tables III and IV). The comparison between Igf2 mϩ/pϪ and K10s⌬Igf2r/ϩ transgenics (Kipps line) with the genetically manipulated element in different mice revealed that the wet weights of organs in the alimentary canal and liver did not differ significantly in the two groups (column B versus C, Tables III and IV). Body weight and wet weights of the skin, kidneys, and heart were significantly higher in Kipps than Igf2 mϩ/pϪ (column B versus column C, Tables III and IV).
Igf2 mϩ/pϪ mice expressing the K10s⌬Igf2r transgene (Igf2 mϩ/pϪ /Kipps and Igf2 mϩ/pϪ /Krishna) showed a body weight similar to Igf2 mϩ/pϪ mice and significantly lower than that of wild type or Kipps (Table III). Among the organs expressing the K10s⌬Igf2r transgene, the wet and dry weights of cecum and colon did not significantly differ in any of the com-parisons involving the following four groups: Igf2 mϩ/pϪ mice, Kipps single transgenics, and the double transgenics Igf2 mϩ/pϪ /Kipps or Igf2 mϩ/pϪ /Krishna (columns B-D, Table  III). By contrast, the wet weights of the stomach and skin were further decreased in Igf2 mϩ/pϪ /Kipps mice compared with Igf2 mϩ/pϪ and so was the dry weight of the stomach (column D versus column B, Table III). Consistent with the relative levels of K10s⌬Igf2r transgene expression, the decrease in the stomach and skin sizes were more pronounced in Igf2 mϩ/pϪ /Kipps than in Igf2 mϩ/pϪ /Krishna double transgenics (column D in Table III (I) and K10s⌬Igf2r/ϩ (K). Comparisons were by paired t test of sex, litter, and age matched mice. For a given organ, r is the ratio between the percentage of variation in body weight and the percentage of variation in organ wet weight, both compared with wild type. Values of r Ͼ 1 indicate that organ size is greater than predicted from body weight. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; no symbol, not significant. lost before the age of 90 days and wild type or K10s⌬Igf2r/ϩ heterozygotes were fully viable.

DISCUSSION
The membrane form of IGF2R binds IGF-II, retinoic acid, and molecules containing mannose 6-phosphate residues (9 -14). This latter group of ligands includes molecules as different in function and structure as the latent form of TGF-␤1 and lysosomal enzymes and possibly other still uncharacterized factors (9). Although IGF-II is the only identified ligand of the soluble IGF2R in vivo, it is likely that the two forms of IGF2R share a similar ligand specificity (19). Because of the potential complexity of its binding activity, the biological functions of sIGF2R are difficult to interpret.
We used a genetic approach to discover whether the soluble IGF2R reduces organ size by interacting exclusively with IGF-II. The combined analysis of mice originated by crossing transgenics expressing a soluble IGF2R by mice overexpressing IGF-II or carrying a disrupted paternal (active) allele of the Igf2 gene allowed the identification of IGF-II-mediated and IGF-II-independent activities of sIGF2R in vivo.
IGF-II, the Soluble IGF2R and the Determination of Adult Organ Size-If all the actions of the soluble IGF2R on growth are solely due to its interaction with IGF-II, then the presence of excess s⌬IGF2R should not change the phenotype of mice that lack IGF-II (Igf2 mϩ/pϪ ). This test was applied to the phenotype.
Igf2 mϩ/pϪ mice retain a dwarf phenotype throughout adult life, long after the developmentally regulated and sharp drop in IGF-II levels that occurs at 20 -40 days after birth (23). This implies that fetal IGF-II levels have long term effects on the size of adult organs. Our results show that the endogenous IGF-II contributes to about 30% of the size of the body. The reduction in body weight of these mice was not as great as the 35-40% reported previously (8,23). This may reflect the mixed strain genetic background used in the present work, because we have observed a similar effect in comparisons of F 1 hybrids with pure strain mice bearing the mutation. 3 We failed to detect a change in DNA content in comparisons of Igf2 mϩ/pϪ mice with wild type. Mice in which signaling of IGF-II is reduced by disruption of the type 1 IGF receptor show hypoplasia in some tissues. In the same mice, cellular density in the spinal cord is higher than in wild type, and no obvious histological abnormality is detectable in the skin of Igf2 mϩ/pϪ embryos (26). The wet weight of most organs in Igf2 mϩ/pϪ mice was reduced at 90 days (Tables III and IV), and these include the cecum, colon, heart, liver, and kidneys. We previously showed that the overexpression of a soluble IGF2R (K10s⌬Igf2r transgene, Kipps and Krishna lines) causes a local reduction in size of the alimentary canal and the uterus. The temporal pattern of expression of the K10s⌬Igf2r transgene is dictated by the ker-atin 10 promoter, and it partly overlaps with that of the Igf2 gene from at least E12.5 on (21). The overexpression of the K10s⌬Igf2r transgene in the colon and cecum in a Igf2 mϩ/pϪ background does not result in any further decrease in the wet weight of the two organs. Furthermore, the wet weights of cecum and colon are not significantly different in Igf2 mϩ/pϪ and Kipps heterozygotes (Table II and Fig. 2). We conclude that the soluble IGF2R acts exclusively by reducing the bioavailability of IGF-II in these organs.
The second test of the interaction between soluble IGF2R and IGF-II employed the Blast transgenic line, which expressed extra IGF-II in a range of organs similar to those where s⌬IGF2R is abundant in the transgenes (22). At the time of the analysis, the Blast transgene had been bred for over 10 generations onto the 129J/Sv background, and its phenotype differed from the original description in two respects. First, the wet weight increase in the alimentary canal was only mirrored in an increased DNA content in the colon (column B in Table I compared with Ref. 22). Second, there was no change in the fat content of the fat pad corresponding to the IV mammary gland (column B in Table II Table I, and Figs. 1 and 2). These data are consistent with the notion that s⌬IGF2R can lower the levels of bioactive IGF-II. Overall, the contribution of the Kipps or Krishna line to the degree of phenotypical changes observed in the crosses discussed in this study correlated with the relative levels of expression of the K10s⌬Igf2r transgene in the two lines (21).
Determination of the Size of the Stomach-The wet weight of the stomach was decreased in Igf2 mϩ/pϪ mice by 16% compared with the 30% decrease observed for other organs and the whole body, suggesting that fetal IGF-II has a relatively minor role in the determination of the size of the stomach in the adult (Table III). This conclusion is consistent with the observation that the stomach is relatively unresponsive to the overexpression of IGF-II, despite the high levels of transgenic IGF-II mRNA detected in this organ (Ref. 22, Table I, and Fig. 2).
The expression of the K10s⌬Igf2r transgene in a Igf2 mϩ/pϪ background resulted in a further decrease in the wet weight of the stomach compared with Igf2 mϩ/pϪ (Table III and Fig. 2). The implication of this result is that s⌬IGF2R controls the size of the stomach by modulating additional factor(s) that are IGF-II-independent. The identity of these factors is uncertain. The only ligands of the membrane form of IGF2R with an established growth regulation activity besides IGF-II are the latent form of TGF-␤1 and retinoic acid (1,13). It is not known whether sIGF2R binds any of the two ligands and whether it is involved in the activation of the latent TGF-␤1. The high serum levels of latent TGF-␤1 in K10s⌬Igf2r transgenics suggest that the soluble IGF2R may act as a reservoir of the growth factor that would then be available for local activation and subsequent growth inhibition (21). The membrane form of IGF2R binds and mediates the angiogenic activity of proliferin (28). sIGF2R may compete for proliferin binding to the cellular form of IGF2R. The soluble receptor may also affect organ size by modulating the activity of mannose 6-phosphate containing enzymes that are involved in the turnover of the extracellular matrix (10,11).
Surprisingly, the wet weight of the stomach in Igf2 mϩ/pϪ / Kipps mice is significantly lower than in Igf2 mϩ/pϪ or Kipps 3 C. F. Graham, unpublished observations. heterozygotes, whereas the two latter groups are indistinguishable (Table III). This observation suggests that receptor molecules that are void of IGF-II can interact with other factors more efficiently. In vitro experiments indicate that the level of occupancy of the binding site for IGF-II can modulate the binding activity of the mannose 6-phosphate binding site in the membrane form of IGF2R (29,30). A similar mechanism may function in the sIGF2R molecule.
Igf2 mϩ/pϪ /Kipps mice suffered high mortality (63%) during the first 4 weeks after birth. A possible explanation is that the greatly reduced size of the stomach (57% the size of wild type) resulted in competitive disadvantage in the litter, leading to malnutrition and death. Postnatal mortality was lower (19%) in Igf2 mϩ/pϪ /Krishna mice, consistent with the less dramatic decrease in stomach size (65% of wild type).
Determination of the Size of the Skin and Uterus-The wet weight of the skin was slightly disproportionate to the rest of the body in Igf2 mϩ/pϪ mice compared with wild type (22% wet weight decrease versus 31% body weight decrease, Table III). Skin weight was increased by 60% following IGF-II overexpression (Table I). Surprisingly, the wet weight of the skin was reduced by the expression of s⌬IGF2R in both Blast and Igf2 mϩ/pϪ backgrounds, although it was unaffected in K10s⌬Igf2r/ϩ transgenics (Tables I and III). During the analysis of the phenotype of K10s⌬Igf2r/ϩ transgenics, we speculated that the lack of responsiveness of the skin to s⌬IGF2R was due to a complex compensatory mechanism preserving the size of the organ (21). Indeed, the data shown in the present study suggest that the size of the skin is controlled by IGF-II (wet weight decrease in Igf2 mϩ/pϪ ; megaly in Blast and attenuation in Blast/Kipps) acting in cooperation with a factor modulated by sIGF2R (reduction in Igf2 mϩ/pϪ /Kipps compared with Igf2 mϩ/pϪ ). The inability of s⌬IGF2R to attenuate the overgrowth of the uterus in Blast is consistent with the low level of expression of the K10s⌬Igf2r transgene in this organ (21).
Organ Parameters Affected by IGF-II and sIGF2R-The total dry weight and water content changed proportionally to the wet weight and to each other in Igf2 mϩ/pϪ mice and in organs overexpressing IGF-II or a soluble IGF2R (21). Our result also indicate that in most organs a significant change in dry weight and therefore total water content is accompanied by a insignificant change in total DNA or detergent-soluble protein contents. Taken together, our data suggest that sIGF2R may act on a set of targets with similar biological activity, one of which is IGF-II.
Another implication is that a major consequence of the alteration of IGF-II levels is a change in fluid and insoluble protein contents. Our data do not unequivocably support the idea that IGF-II causes edema (6 -8, 22). The change in fluid content may be secondary to the changes in dry matter and may not involve intercellular fluids. An accumulation of fluid following an increase in the levels of IGF-II may reflect increased vascular permeability and/or change in the vascularization rate with subsequent increased delivery of nutrients from the circulation. The elements that determine what is the "right" organ size are not well characterized. The observation that the growth of the prostate is tightly regulated by the vascular epithelium points to an interplay between paracrine factors (IGFs and others) and angiogenic factors in determining tissue mass (31,32). The membrane and soluble forms of IGF2R, IGF-II, and angiogenic factors such as proliferin may be elements of a system controlling organ size by dictating the rate of angiogenesis. Also, a mechanism must exist by which fetal IGF-II sets the rate at which fluid and insoluble matter are accumulated by tissues and by which this rate is maintained after the levels of circulating IGF-II decline in the adult.
The general conclusions of this work are: 1) the local expres-FIG. 2. Synopsis of changes in wet weight of the stomach, colon, and kidneys in mice with altered levels of soluble IGF2R and IGF-II. Differences in comparisons with wild type (percentage of reduction or increase) are shown only when statistically significant. Absence of growth factors or soluble IGF2R symbols indicates absence of expression and not necessarily absence of circulating factor produced by expressing organs. In the stomach, the soluble IGF2R must decrease organ size by interacting with factor(s) other than IGF-II, because the wet weight is further decreased in Kipps/Igf2 mϩ/pϪ mice compared with Igf2 mϩ/pϪ . IGF-II buffers the interaction between s⌬IGF2R and the unknown factor(s). Moreover, the overall response of the stomach to IGF-II overexpression (Blast) is limited. In the colon, lack of IGF-II by gene disruption or expression of s⌬IGF2R produce similar decreases in size. The wet weight of the colon is not further decreased by expression of the soluble IGF2R in mice that lack IGF-II. The colon responds to IGF-II overexpression by doubling in size. In the kidneys, the pattern of the changes in wet weight is consistent with the local activity of both Blast and Kipps transgenes. sion of a soluble IGF2R reduces the organomegaly induced by excess IGF-II. 2) This reduction involves wet and dry weight in the alimentary canal and skin, but the DNA content is rarely lowered.
3) The soluble IGF2R does not alter the size of the cecum and colon when mice lack IGF-II. This suggests that the soluble IGF2R can reduce organ size through interactions with IGF-II in mice with normal IGF-II levels. 4) The soluble IGF2R reduces the size of the stomach when mice lack IGF-II. This suggests that the soluble IGF2R can also act through an IGF-II-independent pathway (Fig. 2). Further work will be necessary to determine the molecular bases of IGF-II-mediated and -independent activities of sIGF2R and to identify the primary response to IGF-II in the determination of organ size.