Main

IGF-I and IGF-II have been shown to be important in regulating cellular growth and differentiation(1, 2). Both IGF-I and IGF-II are detectable in human fetal tissues, the amount of IGF-II exceeding greatly that of IGF-I(3–5). IGF-II is believed to be an important regulator of fetal growth and development because of its abundance in multiple fetal tissues as demonstrated by in situ hybridization and immunolocalization studies as well as its high concentration in fetal serum, which is 6-10 times higher than that of IGF-I(6–12). In mice disruption of one of the IGF-II alleles by gene targeting and germ line transmission of the inactivated IGF-II gene from male chimeras yielded heterozygous progeny that were smaller than their wild-type littermates, but continued to grow at a normal rate after birth. This also indicates an important role for IGF-II in fetal growth(13).

In situ hybridization studies revealed that IGF-II mRNA is primarily localized in interstitial cells of the developing human fetus(3), implying that the action of IGF-II may be paracrine in nature rather than acting as an autocrine growth factor. However, hitherto only radioactive probes have been used as detecting systems for IGF-II. Although these techniques seem to be highly sensitive and reproducible, a number of problems exist, i.e. poor structural resolution, instability, and high costs of radiolabeled probes. The reason why in situ hybridization with nonradioactive probes have been less frequently used so far is the belief that they are relatively insensitive and/or tend to produce high levels of nonspecific background(14). With the improved methodology now available, many of these concerns have been alleviated(15). In situ hybridization studies designed to determine the anatomical topology of IGF-II mRNA in fetal tissues may have missed subtle changes in intratissue distribution. Furthermore, no information on the expression of IGF-II protein was provided. Determining more exactly the sites of IGF-II production in fetal tissues is necessary in further delineating the role of IGF-II in fetal organogenesis. The objective of this study was therefore to investigate IGF-II expression in human fetal tissues by in situ hybridization using a digoxigenin-labeled riboprobe and by immunocytochemistry to define in detail the cellular source of IGF-II in human organogenesis.

METHODS

Tissue preparation, pretreatment of sections. Human fetal tissues were collected from prostaglandin-induced abortuses of 18-26-wk gestation with prior approval of the Ethics Committee of the University of Vienna. Cubes of tissue (approximately 1 cm3) taken from brain, liver, lung, kidney, adrenals, muscle, costal cartilage, and placenta were removed and fixed in 4% paraformaldehyde. After fixation, the tissues were washed in PBS and embedded in paraffin. Slides were precoated with 2% 3-aminopropyltriethoxysilane (Sigma Chemical Co., Vienna) and air-dried. After dewaxing, the sections were postfixed in 4% paraformaldehyde in 0.1 M phosphate buffer and treated with 0.2 M HCl. Nonspecific staining was prevented by acetylation. Proteinase K digestion was performed at 37°C in 0.5 M Tris-HCl buffer (pH 7.4). Different concentrations and digestion times were tested. The enzymatic reaction was stopped with Tris-HCl buffer at 4°C. Then the sections were dehydrated in graded ethanol and chloroform.

Probes and reagents. The IGF-II probe used was a single-stranded cRNA. Sense and antisense cRNA probes were transcribed from full-length cDNA fragments subcloned into pCEM 72 (Szabo, Vienna) using either SP6 or T7 polymerase; sense probes were used as controls. RNA labeling was performed by in vitro transcription with T3 and SP6 polymerases. Before use the labeled probes were purified twice over a Sephadex G50 column(Pharmacia, Freiburg, FRG) to completely remove unincorporated digoxigenin-labeled nucleotides. The efficiency and intensity of labeling was monitored by dot-blots. Dots (1 μL) of the probe at serial dilutions were spotted on a nylon membrane, dried, and UV-linked for 2 min. Unsaturated binding sites were blocked with blocking reagents (Boehringer Mannheim, FRG). Digoxigenin labeling was detected with alkaline phosphatase-conjugated anti-digoxigenin, as previously described(15).

Reagents were purchased from the following sources: Boehringer Mannheim, FRG: Dig RNA labeling kit, antidigoxigenin AP, Fab fragments, 4-nitro blue tetrazolium chloride, 5-bromo-4-chloro-3-indolyl phosphate (X-phosphate); Sigma Chemical Co., St. Louis, MO: proteinase K (P0390), DNA (D1626), 3-aminopropyltriethoxysilane (A3648), N,N-dimethylformamide (D8654); and Merck, Darmstadt, FRG: paraformaldehyde, acetic anhydride, formamide, dextran sulfate.

Hybridization. In situ hybridization was performed on paraffin-embedded tissues as previously described. Briefly the probes were denatured, and the probe solution was pipetted onto the slides, then covered with a coverglass and placed on a hot plate at 95°C for 4 min. The hybridization solution contained: 50% deionized formamide, 2 × SSC, 10% dextran sulfate, 0.01% sheared salmon sperm DNA, 0.02% SDS, and the probe in appropriate concentration up to 20%, depending on the labeling intensity, as determined in dot-blots. Hybridization was performed at 37 or 55°C for 4-16 h. After hybridization, the sections were washed for a minimum of 2 h in 2 × SSC. Stringency conditions were calculated from melting temperatures of individual probes. Digoxigenin labeling was detected with alkaline phosphatase-conjugated anti-digoxigenin. Development of the in situ hybridization was performed in 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl phosphate.

Sense cRNA IGF-II was used as control for the detection of IGF-II mRNA. To test for nonspecific binding of the secondary detection systems, hybridization was also performed in the absence of specific probes.

Immunocytochemistry. Immunocytochemistry was performed with a biotin-avidin or an alkaline phosphatase/antialkaline phosphatase technique as described in detail previously(16). We used a commercial MAb against IGF-II derived from rat spleen cells purified by ammonium sulfate precipitation and diethylaminoethyl-cellulose chromatography at a dilution of 1:250 (Biomedica). Cross-reactivity of this antibody with IGF-I had been shown previously to be less than 5%(17).

RESULTS

We have screened seven fetal tissues and the placenta for IGF-II mRNA by in situ hybridization and immunocytochemistry. The distribution of IGF-II in human fetal tissues as determined by in situ hybridization and immunocytochemistry is summarized in Table 1.

Table 1 Cellular localization of IGF-II in human fetal tissues
Full size table

In costal cartilage strong IGF-II expression was present in perichondrial cells and the stellate-shaped primitive chondroblasts, whereas ovoid and rounded chondroblasts exhibited a weaker positive reaction; immunohistochemistry showed identical intensity and distribution (Fig. 1,a and b).

Figure 1
figure 1

Cartilage: positive staining of immature spindled chondroblasts in both in situ hybridization (a) and immunohistochemistry (b). Liver, strong positivity for IGF-II mRNA(c) in hepatocytes and hepatic artery, negativity in bile duct(large arrow), hematopoietic cells, and endothelium of the portal vein (small arrow). Immunohistochemical localization of IGF-II(d), positivity of hepatocytes, negatively staining endothelium of the central vein (arrow). Kidney, IGF-II mRNA expression(e) in blastemal cells of the nephrogenic zone and in the primitive epithelium (arrow). Immunohistochemical localization of IGF-II(f) showing most constant positivity in developing proximal tubule close to the glomerulus. Muscle (g), distinct positivity of myocytes with accentuating perinuclear region and cross striation. Adrenal gland(h), cellular positivity in the persistent zone and in the sinusoidal endothelium and interstitial cells of the fetal zone. Placenta(i), negative IGF-II mRNA staining for syncytiotrophoblast(arrow), positivity in underlying cytotrophoblast (filled arrow), endothelial and interstitial cells. Magnifications: a, b, and g, ×800; c-f, h, and i, ×400.

Full size image

In the liver the IGF-II probe hybridized strongly to hepatocytes, whereas only occasional hematopoietic cells with large nuclei contained positive signals. Endothelial cells of the hepatic artery showed a positive reaction; however, both Kupffer cells and bile ducts in the central position of the portal tracts exhibited no reaction. Immunohistochemical reactions for the detection of the IGF-II protein were in concordance with the results of the in situ hybridization; hepatocytes and endothelial cells of the hepatic artery showed distinct positivity and Kupffer cells and bile ducts showed no reaction (Fig. 1,c and d).

The kidneys of second trimester human fetuses are relatively well differentiated. Although early epithelial differentiation of the mesenchymal blastema still occurs in the outer cortical areas, the inner regions contain well developed metanephric blastema-derived excretory units (glomeruli and proximal and distal convoluted tubules with the intervening loops of Henle) and collecting tubules from the ureteric bud. IGF-II mRNA was expressed in the blastemal cells of the nephrogenic zone and in interstitial cells of the medullary region. Furthermore, positivity was observed focally in the primitive epithelium of the S-loop and within the ureteric epithelium. No IGF-II mRNA was present in differentiated tubular segments of the nephron. By using immunohistochemistry we observed weak positivity in blastemal cells but strong positivity in developing proximal tubules, accentuating the first postglomerular segment (Fig. 1,e and f).

Skeletal muscle had distinct perinuclear positivity in the fused myocytes of the developing part of the muscle. In addition, a positive signal was observed in the interstitial cells between the fibers (Fig. 1g). These findings were confirmed by immunohistochemistry; interstitial cells and myocytes showed strong positivity for IGF-II.

Adrenal epithelium exhibited no positive reaction for IGF-II mRNA in the centrally located fetal zone of the cortex, whereas the peripheral,i.e. persistent, zone showed very strong IGF-II expression as did the sinusoidal endothelium in all parts of the fetal cortex (Fig. 1h). Immunohistochemical analysis showed only poor reactivity for IGF-II in the peripheral zone of the fetal adrenal gland; endothelial and interstitial cells were positive.

In the placenta stromal cells of placental villi showed IGF-II mRNA expression, which was stronger and more frequently noted in peripheral villi. Endothelial cells of peripheral capillaries and central arteries also showed a prominent positivity. An intense IGF-II labeling was observed in decidual cells and was slightly less prominent within the proliferative cells of the cytotrophoblast. The superficial layer of syncytiotrophoblast lacked detectable IGF-II mRNA (Fig. 1i). Immunoohistochemistry revealed positive reactions for endothelial cells, interstitial cells, and in the cytotrophoblast; however, the positivity observed was slightly less intense than in the in situ hybridization.

In the human fetal cerebrum all stages of neuroepithelial cell development and differentiation could be assigned to distinct zones as early as 16 wk of gestation(18). Thus, despite the limited availability of human fetal material, the tissue we studied representatively covered fetal brain development. Epithelial cells of the choroid plexus showed marked IGF-II mRNA expression; endothelial cells also exhibited positive staining(Table 1, Fig. 1, a-i). No IGF-II mRNA expression was present in both neuroblasts of the germinal matrix and mature ganglionic cells, nor was reactivity seen in glial cells; immunohistochemical results were in accordance with those of in situ hybridization.

In the lung strong positivity for IGF-II mRNA was detected within the developing bronchial cartilage and the endothelial lining of intraparenchymatous pulmonary arteries. Positive reaction of the pulmonary epithelium appeared to increase with duration of pregnancy and could be seen quite commonly in medium-sized bronchiolar tubules of the canalicular stage in fetuses of 22-26 wk of gestation, whereas in the early alveolar stage it was restricted to single cells of bronchial and bronchiolar linings. Immunohistochemistry showed less intense positivity for IGF-II in pulmonary epithelial cells, identical intensity was found in bronchial cartilage and interstitial cells.

Endothelial IGF-II mRNA positivity was a constant finding in all organs, but varied according to size and type of the vessels. The most intense labeling was noted in medium-sized arteries and arterioles as well as in the sinuses of liver and adrenal gland. In contrast, with the exception of some larger hepatic veins, no reaction was visible in venous vessels. Interstitial cells of all organs showed a distinct positive reaction, and the most intense labeling was observed in the renal medulla of the immature kidneys.

DISCUSSION

To elucidate the specific cellular sources and targets of local IGF-II produced in the developing fetus, we used in situ hybridization and immunocytochemistry to map the cellular patterns of IGF-II expression. In our investigation of human fetal tissues of 18-26 wk of gestation IGF-II mRNA was detectable in all tissues studied. However, in comparison with the findings of Han et al.(3), the pattern of IGF-II mRNA expression differed in a variety of ways. This may well be due to the different probe we applied in our study, i.e. a digoxigenin-labeled nucleotide with a greater number of base pairs compared with the short oligonucleotide probes used by Han. Hirvonen et al.(18), too, noticed a greater sensitivity using a nearly full-length cDNA probe; however, the precise cellular location of IGF-II gene expression was not demonstrated by these authors. Owing to the high structural resolution of our method, it was possible to locate IGF-II mRNA specifically in distinct cell types(15).

We found a strong positivity for IGF-II mRNA in hepatocytes of all fetuses between 18 and 26 wk of gestation with the digoxigenin-labeled probe, demonstrating that the ISH method used by us was more sensitive in detecting IGF-II mRNA, in, e.g. fetal hepatocytes, than was the method using radioactive probes described in Han et al.(3).

IGF-II mRNA is heavily expressed in the fetal kidneys(18), its expression is continued through adulthood, and has been attributed an autocrine and/or paracrine role in nephrogenesis. However, IGF-II may not be essential for normal renal development as transgenic mice, lacking the IGF-II gene, have normal renal morphology and function(19). IGF-I and IGF-II expression patterns are significantly different in rodents and in the human: IGF-II mRNA levels are very high in in both rodent and human fetal kidneys(18, 20) and remain fairly abundant in the adult human kidney(20). This has led to the suggestion that IGF-II is involved in human renal development(21). In our investigation IGF-II was highly abundant in the nephrogenic zone of the fetal kidney in addition to IGF-II expression in the vascular system, which differs from IGF-II expression pattern in the adult kidney(20), where it is concentrated in renal vascular system only. The function of renovascular IGF-II may consist of a local trophic role for the vascular system, or it may enter the blood stream and contribute to the circulating pool of IGF-II. Our results show that IGF-II, in addition to its high degree of expression in the vasculature, is markedly expressed in the developing part of the kidney.

Using Northern blot analysis IGF-II has recently been shown to be present in high abundance in human fetal adrenal glands and to be regulated by ACTH in cultured adrenal cortical cells(22). This is inconsistent with previous findings that IGF-II mRNA is expressed only in mesenchymal cells of the capsule during midgestation and in the cortex only earlier in gestation (18 d to 14 wk). However, our data support the observation that IGF-II mRNA is expressed in both the definite and fetal zone of the cortex; this pattern of hybridization was seen in adrenals from the 18th to the 26th wk of gestation and did not seem to change during this period of gestation. Therefore, it seems that IGF-II is a significant growth factor for the human fetal adrenal gland.

Our results are not consistent with the previously published observation of Han et al.(3) showing that IGF-II is expressed predominantly in interstitial cells of the organs investigated. In general, tissues of mesodermal origin express more IGF-II than those originating from the other two germ layers(23–25). Our studies, however, clearly demonstrate localization of IGF-II mRNA in hepatocytes, adrenal cortical cells, and ureteric epithelium and suggest that IGF-II may also have an autocrine action.

IGF-II expression is mandatory for myoblast differentiation in myoblast cell lines(26, 27). Listrat et al.(23), using a 35P-labeled IGF-II cRNA probe, localized the majority of IGF-II transcripts in bovine fetuses to developing muscle rather than to connective tissue. This is similar to the expression pattern shown in developing skeletal muscle of rat embryos(28). Connective tissue seems to play an active role in modulating myogenesis possibly by sequestering IGF-II through a tissue- and time-dependent production of IGFBPs(23, 29, 30). These observations are in line with our results showing a distinct IGF-II mRNA expression in myocytes and interstitial cells.

The widespread distribution of IGF-II mRNA underlines its crucial role in fetal development. Its distribution pattern is similar to that of IGFBP mRNAs, which implies that there is a coordinated expression of the IGF and IGFBP genes(5, 31). This suggests that the binding proteins and the IGFs are produced in close proximity and that their interaction may regulate the effect of the IGFs on their receptors near the sites of IGF production(31, 32). This type of coordinate expression seems to be also important in the appropriate timing of cellular events such as differentiation, as has been demonstrated in muscle development(33–35) and is likely to be true for other fetal tissues as well.

The lack of concordance among some cells identified by the two techniques suggests that some of the cells identified by immunohistochemistry do not synthesize IGF-II. Furthermore the IGF-II mRNA of some fetal cells may be unstable, not translated, or result in production of a peptide form that is not recognized by the antibody used in our study. The MAb used in our study has <5% cross-reactivity with IGF-I, which is markedly lower than the approximately 50% reported in other studies(36). Some immunostained cells may have sequestered IGF-II from other sources by cell surface binding with subsequent internalization and sequestration in endoplasmic reticulum. Interaction of ligands of the IGF-II/M6P receptor, interference of cellular uptake by steric hindrance, or conformational changes of the receptor molecule may have caused inhibition of IGF-II uptake and subsequent differences in detection(37).

Due to the involvement of the IGF system in fetal growth and development, differences in the expression of the IGF system may be in line with the developmental and maturational pattern in organogenesis and embryogenesis. Furthermore, reduction of placental size and blood flow restricts fetal growth to a greater extent than either nutrient or oxygen deficit alone. Anatomical or functional placental insufficiency may be caused by alterations in IGF-II expression, because on the one hand the high expression of IGF-II in decidual cells suggests an interaction between trophoblast and decidua, on the other hand IGF-II may exert a local trophic role on the vascular system and be important for modulating growth of adjacent cells.

In conclusion, our results suggest that the distribution of IGF-II in human fetal tissue is much more widespread than hitherto thought and can be localized specifically. We have shown that the digoxigenin-labeled IGF-II probe that we have used is very sensitive for the detection of IGF-II mRNA expression in human fetal tissues(15) and is suitable for analysis of routinely prepared paraffin-embedded material, when section pretreatment, hybridization conditions, and signal amplification are optimized.