Testicular atrophy was induced as previously described 10. Briefly, male Wistar rats (4 weeks old, 150–160 g) were subjected to 1.2 mg/Kg b.w. five times/week intra-scrotal injections of epinephrine in sterile saline (Sigma), for 11 weeks. Rats were housed in cages placed in a room with 12-hour light-dark cycle and constant humidity and temperature (20°C). Both, food (standard semipurified diet for rodents; B.K. Universal, Sant Vicent del Horts, Spain) and water were given ad libitum. Healthy age-matched control rats were studied in parallel. All experimental procedures were performed in conformity with The Guiding Principles for Research Involving Animals.

After testicular atrophy induction (after 11 weeks receiving epinephrine), rats with testicular atrophy were randomly assigned to receive either vehicle (saline) (Group AT, n = 10) or recombinant human IGF-I (Chiron) (2 μg × 100 g bw^-1xday ^-1 in two divided doses, subcutaneously) (Group AT+IGF, n = 10) for four weeks. Control rats (Group CO, n = 10) received saline during the same period.

In the morning of day 0 (before treatment), animals were weighted and blood samples were drawn from the retroocular venous plexus from all rats with capillary tubes (Marienfeld, Germany). Serum samples were stored at -20°C until used for analytical purposes.

In the morning of the 29^th day, rats were weighted, blood was obtained again and processed as previously indicated and animals were sacrificed by decapitation. After the abdominal cavity was opened, the testes were dissected and weighted. The testicular diameters (AP and LM) were measured in testes previously fixed for histology, using a precision calliper, Mituyoto^® (± 0.05 mm).

Bouin-fixed tissues were processed and sections (4 μm) were stained with Haemotoxylin and Eosin and Masson's trichrome. For histopathological evaluation of testes 30 seminiferous tubules from each rat of the three groups were blindly evaluated by two observers and the arithmetic mean of the scores was taken as the final result. Transversal sections of seminiferous tubuli were examined and evaluation of histological changes was made using a light projection microscope (Micro Promar Leitz GMBH, Wetzlar, Germany) at 100 × magnification. The following parameters were studied: tubular diameter, quantitation of the presence of the different types of cells in tubuli, presence of peritubular fibrosis, and the number of proliferating cells. For general purposes Haematoxilin & Eosin and Masson's trichrome staining were used. Specific techniques for other purposes are specified in the corresponding paragraphs.

Tubular diameters were expressed in μm. Changes in tubuli were classified into five categories (Category I: highest damage to Category V: full normality). Category I: presence of only Sertoli cells; Category II: Sertoli cells plus spermatids, Category III: Sertoli cells, plus spermatides, plus spermatocytes, Category IV: presence of all kinds of cells but showing some morphological alterations (i.e.: severe vacuolization, aberrant cells). Category V: presence of all kinds of cells without morphological alterations. The presence of peritubular fibrosis was evaluated in Masson's trichrome preparations according to the thickness of the staining of collagen deposition surrounding tubuli. Proliferating cells were identified by immunostaining of proliferating cellular nuclear antigen (PCNA) using an avidin-biotin peroxidase method 11 with retrieval of antigen by means of microware irradiation. Specific anti PCNA antibody (mouse anti-PCNA, clone PC 10, DAKO, Denmark) biotinylated rabbit anti-mouse IgG (DAKO, Denmark) were used and the avidin-biotin complex technique (ABC, DAKO kit) was performed. The bound antibodies were visualized by means of 3,3'-diaminobenzidine tetrahydrochloride (SIGMA Chemical Company, St. Louis, MO) with nickel enhancement (Shu et al., 1988) 11. Finally, samples were slightly counterstained (10 seconds) in hematoxilin, dehydrated, and mounted in DPX. Controls were performed by substitution of the primary antibody by TBS. The number of PCNA positive cells was recorded. The result was expressed as stained cells per tubuli (arithmetic mean of 30 screened tubuli).

Taking into account all the parameters specified above, an overall score of testicular histopathological damage was adopted according to the following guidelines: Tubular diameter (in μm) scored from 0 to 3 points: > 260 = 0 points; from 240 to 259 = 1 point; from 220 to 239 = 2 points; and < 219 = 3 points. Cellular counts in tubuli: Category I (8 points), category II (6 points), category III (4 points), category IV (2 points) and category V (0 points). The score was obtained by multiplying the number of tubuli in each category by its respective points divided by 30 (the number of tubuli evaluated in each animal). Peritubular fibrosis was scored from 0 (absent or minimal) to 1 (evident). Cellular proliferation (PCNA) was scored from 0 to 3 according to the following criteria: when the number of PCNA positive cells /tubule was higher than 60, the score was 0 point. When the number of PCNA positive cells / tubuli was lower than 60, the score was obtained according to the following formula: (60 – PCNA positive cells) × 0.05. Therefore the overall score of histopathological damage ranged from 0 (complete normality) to 15 (full abnormality).

In addition, the expression of transferrin in tubuli was evaluated by immunostaining using similar technique as for PCNA with specific anti-transferrin antibody (obtained from rabbit, RARa/TRf, Nordic Immunological Laboratories, Teknovas, The Netherlands). Transferrin expression was scored from 0 to 4 points. If 30 tubuli expressed transferrin normally all over the germinal epithelium: 0 points. The remaining scores were obtained according to the following formula: (30 - Tubuli showing expression of transferrin all over the germinal epithelium) × 0.075.

Serum levels of albumin, total proteins, glucose, cholesterol and alkaline phosphatase, were determined by routine laboratory methods using a Hitachi 747 autoanalyzer (Boerhringer-Mannheim, Germany).

Serum levels of the different hormones were assessed by RIA in a GammaChen 9612 Plus (Serono Diagnostics, Roma, Italy) using specific commercial assay systems. The sensitivity (S) of total Testosterone assay was 4 ng/dL and the intraassay coefficient of variation (CV) was less than 7%. The sensitivity of free Testosterone assay was 0.15 pg/mL and the CV was <8%. The sensitivity of Estradiol was 7 pg/mL and the CV was <7% (Coat-a-Count, DPC (Diagnostic Products Corporation, Los Angeles, CA). The kits for Rat luteinizing Hormone (rLH) (S = 1.7 ng/mL and CV<10%) and Rat follicle stimulating hormone (rFSH) (S = 1.8 ng/mL and CV<6%) were provided from Amersham International plc (Little Chalfont Buckinghamshire, England HP7 9NA). Assessment of IGF-I was carried out after an alcohol extraction in order to eliminate IGFBPs (Nichols Institute Diagnostics, San Juan Capistrano, CA, USA). The sensitivity of this assay was 21 ng/mL and the intraassay coefficient of variation was less than 4%.

Antioxidant enzyme activity phospholipid hydroperoxide glutathione peroxidase was determined as previosuly reported 12. Briefly, PHGPx catalizes the oxidation of reduced glutathione (GSH) by phospholipids hydroperoxide in presence of glutathione reductase and NADPH. Oxidized glutathione (GSSG) is immediately reduced to GSH with the oxidation of NADPH to NADP⁺. And the decrease of NADPH absorbance at 340 nm is assessed using a Cobas Mira autoanalyzer (ABXMicro, Germany). Finally, a unit of PHGPx was expressed as the amount of enzime required to oxide 1 μmol/min of NADPH at 37°C subtracting the blank of the reactives (S = 3.7 IU/L and CV<10%).

IGF binding proteins (IGFBPs) were studied using Western blot analysis which was performed as described by Hossenlopp et al 13 and Hardouin et al. 14. Briefly, rat plasma (3 μL) was diluted up to 50 μL in 0.06 M Tris-HCl and 0.15 M NaCl, pH 6.8, with 5% SDS, 20% glycerol and 0.02% bromophenol blue. The solution was submitted to a 5–15% gradient SDS-PAGE. Proteins were blotted to a 0.2 μm nitrocellulose filter (Bio-Rad). Nitrocellulose filter was dried for 5 min at 37°C, followed by quenching firstly in Tris buffer alone, and then in Tris buffer with 3% NP-40 (30 min at 4°C). Finally, the filter was soaked in Tris buffer with 1% BSA and 0.1% Tween-20 for 60 min, and was then incubated over-night with 200000 cpm ¹²⁵I-IGF-I. After washing several times with buffer, the filter was dried and, lastly, exposed to an autoradiographic film.

Densitographic quantitation of IGFBPs was expressed in arbitrary units of optic density (mm²)

Data are expressed as mean (± SEM). To assess the homogeneity among the three groups of rats a Kruskall-Wallis test was used, followed by multiple post-hoc comparisons using the Mann-Whitney U test (two tailed) with Bonferroni adjustment. A regression model was fitted considering histopathological score, PCNA or transferrin expression scores and IGF-I plasma concentration as the dependent and independent variables respectively. Within groups differences between pre- and post-treatment values were assessed by means of Wilcoxon matched pairs signed rank sum test. Any p value less than 0.05 was considered to be statistically significant. Calculations were performed with SPSSWin v.6.0. Program.

There was a reduction in testicular size and volume in group AT as compared with control (Figure 1). Morphometric study showed a significant reduction in testicular weight (more expressive when values were corrected by body weight) in group AT compared to controls and AT+IGF group. Noteworthy, no significant differences were found between AT+IGF and controls. Longitudinal and transversal testicular diameters were reduced in AT group and increased significantly in rats treated with IGF-I (AT+IGF). Morphometric data in the three groups are summarized in Table 2.

In AT group histological examination revealed testicular atrophy with marked decrease in tubular diameter expressed as an index of damage scored from 0 to 3 points (see Methods) (CO: 0.25 ± 0.05; AT: 2.07 ± 0.19, p < 0.001). IGF-I therapy induced a significant improvement of this parameter (1.45 ± 0.18, p < 0.05 vs group AT and p < 0.01 vs controls). Histopathological findings in untreated AT rats included: vacuolization of Sertoli's cells, loss of germinal line, detached germ cells, dramatic reduction of spermatogenesis, presence of abnormal spermatids or empty tubulli, and peritubular fibrosis. These alterations were much less intense or almost absent in rats receiving IGF-I (see Figure 2 for morphological comparison among groups). These changes are presented in Table 3 according to a pre-established score (see Methods), after exploring 30 tubuli from each preparation.

Testicular cellular proliferation, as evaluated by PCNA positive cells, was significantly reduced in both groups with testicular atrophy: IGF-I therapy induced an improvement but it did not reach statistical significance (mean of PCNA + cells scored in 30 tubuli for each animal; CO: 63.8 ± 1; AT: 36 ± 8, p < 0.001 AT vs CO; AT+IGF: 47 ± 5, p < 0.05 CO vs AT+IGF). Figure 3 shows PCNA immuno-histochemistry in the three groups.

Taking together all the morphological data a score of histopathological testicular damage was established (see Methods). The values of histopathological score were: CO: 0.49 ± 0.14; AT: 12.44 ± 1.29 (p < 0.001 AT vs CO); and AT+IGF: 7.78 ± 1.38 (p < 0.001 vs CO and p < 0.01 vs AT), demonstrating severe testicular damage in untreated AT rats and a significant improvement in AT rats treated with IGF-I.

Testicular transferrin, a marker of the integrity of the hemato-testicular barrier and Sertoli cell function, was evaluated from 0 to 4 points by immunohistochemistry in testicular slices (see Methods). Transferrin expression was decreased in AT rats (1.83 ± 0.43 points, p < 0.001 vs groups CO and AT+IGF) as compared with controls (3.85 ± 0.15 points) and to rats with testicular atrophy receiving IGF-I (2.60 ± 0.30 points, p < 0.001 vs groups CO and AT+IGF). A close inverse correlation was found between testicular transferrin expression and histopathological testicular damage score (r = -0.85, p < 0.001).

Untreated AT rats showed a significant reduction of serum levels of both total (p < 0.05) and free testosterone (p < 0.001) with increases of LH and FSH (p < 0.05) levels as well as the estradiol/testosterone ratio. However, in the group of rats with testicular atrophy treated for four weeks with low doses of IGF-I, a partial recovery was observed with an increase of the levels of both free (p < 0.05 vs untreated group) and total testosterone (p = 0.06 vs untreated group) and a significant reduction of the estradiol/total testosterone ratio (Table 4).

Another interesting result was that comparing data from before and after treatment testosterone concentrations were increased (p < 0.05 before and after treatment) in all animals treated with IGF-I (Figure 4).

At the time of sacrifice (day 29^th), 14–16 hours after the last IGF-I injection, surprisingly serum levels of IGF-I did not increase in rats with testicular atrophy treated with IGF-I (AT+IGF) (CO: 2257 ± 183 ng/ml; AT: 2209 ± 91 ng/ml; AT+IGF-I: 1910 ± 76 ng/ml).

The main IGFBPs were determined by Western-ligand-blot, showing changes in the pattern of IGFBPs with a significant increment of IGFBP3 in rats with testicular atrophy treated with IGF-I (Arbitrary Units, CO: 23.77 ± 0.48, AT: 23.01 ± 0.95, AT+IGF: 28.24 ± 1.11) (see Figure 5). An increase of IGFBP 1 and 2 (AU, CO: 3.31 ± 0.68; AT: 2.90 ± 0.78; AT+IGF: 6.47 ± 1.74) was also observed but it did not reach statistical significance.

The values of PHGPx, an antioxidant enzyme, showed a significant increase in the AT+IGF group with regard to untreated animals with testicular atrophy (CO: 6 ± 0.2 IU/mg of protein; AT: 5 ± 0.3 IU/mg of protein; AT+IGF: 8 ± 0.7 IU/mg of protein), showing an antioxidant effect of IGF-I therapy on testes.