Role of insulin-like growth factor binding proteins in human post-nephrectomy proximal tubule cells

Abstract

1. In order to determine the role of the insulin-like growth factor-I (IGF-I)/IGF binding protein (IGFBP) axis in the augmentation of tubule growth and function following reductions in nephron mass, primary cultures of human proximal tubule cells (PTCs) were generated from the histologically normal sections of ten surgically removed kidneys.

2. PTC hypertrophy (cellular protein content), DNA synthesis (thymidine incorporation) and apical sodium-hydrogen exchange (NHE) activity (ethylisopropylamiloride-sensitive apical ²²Na⁺ uptake) were measured following 24 h incubation in media supplemented with 10% pre- or post-nephrectomy sera obtained from these patients. The results were compared with the effects of pre- and post-operative control sera collected from seven patients undergoing retroperitoneal operations not involving removal of renal tissue.

3. Day 1 post-nephrectomy sera promoted a significant 73% increase in apical NHE activity, which was accompanied by a significant increase in PTC binding of ¹²⁵I-IGF-I (post- vs. pre-nephrectomy, 163 ± 6 vs. 142 ± 4 fmol (mg protein)⁻¹; P < 0.05). Subsequent post-nephrectomy sera significantly stimulated PTC protein content and thymidine incorporation, peaking at day 7 (127.7 ± 14.0 and 118.4 ± 9.0% of pre-nephrectomy values, respectively; P < 0.05). The growth effects were cell specific, as they were not observed with renal cortical fibroblasts. No change was detected in any of these measured variables following exposure to control sera.

4. Serum IGF-I and IGFBP-1 levels did not significantly change over time or between groups. IGFBP-3 levels progressively decreased in both control and nephrectomized sera from pre-operative values of 3580 ± 305 and 3360 ± 217 ng ml⁻¹, respectively, to 2670 ± 341 and 2600 ± 347 ng ml⁻¹ at 1 week post-operation. Serum IGFBP-2 levels increased to a comparable extent in both controls (day 0 vs. day 7, 2940 ± 1024 vs. 7010 ± 2520 ng ml⁻¹; P < 0.01) and nephrectomized patients (day 0 vs. day 7, 3070 ± 656 vs. 9130 ± 2010 ng ml⁻¹; P < 0.01).

5. The results indicate that nephrectomy engenders the elaboration of one or more humoral factor(s), which promotes increased binding of IGF-I to PTCs and which may in turn specifically stimulate PTC Na⁺ transport and growth.

Unilateral nephrectomy results in a rapid and specific stimulation of growth and function of the remaining kidney (Ogden, 1967; Fine, 1986). In experimental animals, the bulk of such renal growth is accounted for by hypertrophy of the proximal tubule, which is detectable within 24-48 h and is preceded by an increase in proximal tubule sodium reabsorption via activated apical sodium-hydrogen exchange (NHE) (Fine, 1986; Pollock & Field, 1993).

Although the existence of a kidney-specific humoral growth factor, which incites and/or regulates compensatory renal growth, has been long established by parabiotic experiments (Van Vroonhoven, Soler-Montesinos & Malt, 1972; Austin, Goldin & Preuss, 1981; Malt, 1983), serum injections in live animals (Lowenstein & Stern, 1963; Austin et al. 1981; Malt, 1983; Pollock, Nobes, Gyory, Heng & Field, 1996) and in vitro assays (Austin et al. 1981; Malt, 1983; Yamada, Kanetake, Saito, Kondo & Yamamoto, 1983; Yamamoto, Kanetake & Yamada, 1983; Yun, Areas, Yamamoto & Preuss, 1988; Esbrit, Garcia Ocana, Garcia Canero, Manzano & Jiminez Clavero, 1991; Garcia Ocana & Esbrit, 1994; Nobe, Pollock, Heng & Field, 1995), its precise identity remains elusive. Attempts at characterization have so far suggested that this factor is species specific (Yamamoto et al. 1983; Fine, 1986; Yun et al. 1988), unaffected by dialysis or heating (Fine, 1986), and possibly synthesized by the liver and activated by the remnant kidney in a time-dependent fashion following nephrectomy (Dicker, Morris & Shipolini, 1977; Fine, 1986; Garcia Ocana & Esbrit, 1992).

Insulin-like growth factor-I (IGF-I), a 7.5 kDa peptide produced in the liver and kidney, exerts renotropic effects in proximal tubule cells (PTCs) and enhances Na⁺ reabsorption in in vitro and in vivo studies (Zumkeller & Schofield, 1992; Hammerman & Miller, 1993; Nobes et al. 1995; Feld & Hirschberg, 1996; Johnson et al. 1997b). A humoral role in the development of renal hypertrophy has been postulated since binding of IGF-I to PTCs is enhanced following reductions in renal mass (Polychronakos, Guyda & Posner, 1985; Hise, Lahn, Shao, Mantzouris & Fontana, 1993), endogenous renal cortical IGF-I accumulation precedes compensatory renal growth (Stiles, Sosenko, D'Ercole & Smith, 1985; Fagin & Melmed, 1987; Flyvbjerg, Thorlacius Ussing, Næraa, Ingerslev & Orskov, 1988; Lajara et al. 1989; Zumkeller & Schofield, 1992; Feld & Hirschberg, 1996; Gronboek, Nielsen, Flyvbjerg & Orskov, 1997) and administration of a somatostatin analogue prevents compensatory renal growth and increased renal IGF-I content (Flyvbjerg, Frystyk, Thoracius Ussing & Orskov, 1989; Hammerman & Miller, 1993). Although serum IGF-I levels are reported to be either unchanged (Stiles et al. 1985; Fagin & Melmed, 1987; Lajara et al. 1989; Hammerman & Miller, 1993; Gronboek et al. 1997) or decreased (Flyvbjerg et al. 1988) and renal IGF-I mRNA levels are either unchanged (Lajara et al. 1989) or increased (Fagin & Melmed, 1987) following reductions in renal mass, the accumulation and enhanced action of IGF-I at the proximal tubule following uninephrectomy might be explained by alterations in circulating levels of IGF binding proteins (IGFBPs), which tightly regulate the delivery of serum IGF-I to tissues (Jones & Clemmons, 1995; Feld & Hirschberg, 1996). Circulating IGF-I may then be trapped in the proximal tubule by post-nephrectomy increases in IGF receptor number/affinity or in cell-associated IGFBPs, which potentiate IGF-I action on PTCs (Johnson et al. 1997c). However, the circulating levels of IGFBPs following unilateral nephrectomy and their effects on IGF-I binding to PTCs have not been studied to date.

Therefore, the aim of the present study was to determine whether alterations in the IGF-I/IGFBP axis and its direct interaction with human PTCs account for the renotropic effect of serum following a reduction of renal mass.

METHODS

Patients

Segments of macroscopically and histologically normal renal cortex were obtained aseptically from patients undergoing unilateral nephrectomy for small (< 6 cm) renal adenocarcinoma (n = 5), pelvic transitional cell carcinoma (n = 3), benign complicated renal cyst (n = 1) or angiomyolipoma (n = 1). Sera were also collected from these patients prior to operation (day 0) and 1, 3 and 7 days after nephrectomy.

Control sera were collected at the same peri-operative time points from patients undergoing either pyeloplasty for mild-to-moderate pelviureteric junction obstruction (n = 1), nephrolithotomy for non-obstructive urolithiasis (n = 3), renal exploration for ultimately benign renal lesions (n = 2) and retroperitoneal dissection of benign tissue following successful chemotherapy for testicular carcinoma (n = 1). The selection of this group of subjects represented an attempt to control for the non-specific effects of anaesthesia, retroperitoneal surgery, inpatient stay and post-operative fluid and nutritional management.

All serum samples were collected at approximately 08.00 h following an overnight fast. Operations were performed in the mornings between 08.30 and 11.00 h to minimize diurnal effects. Post-operatively, all patients received intravenous isotonic saline for approximately 48 h and none required blood transfusion.

Exclusion criteria included a pre-operative glomerular filtration rate less than 80 ml min⁻¹ (estimated by the Cockcroft-Gault equation), ureteric obstruction, diabetes mellitus and medications exerting primary or secondary effects on renal function (including diuretics, acetylcholinesterase inhibitors or digoxin). Tissue was not used if operative inspection or histological examination of renal cortex in an area removed from primary pathology revealed interstitial fibrosis or tubular atrophy/hypertrophy. Written informed consent was obtained prior to operation and the collection of blood and use of human renal tissue for primary culture was reviewed and approved by the Royal North Shore Hospital Human Medical Research Ethics Committee.

Cell culture

The method for primary culture of human PTCs and renal cortical fibroblasts (CFs) is described in detail elsewhere (Johnson, Brew, Porronik, Cook, Field & Pollock, 1997a; Johnson et al. 1997c). All experiments were performed on quiescent, confluent, passage 2 PTCs or CFs. Cells were made quiescent by three washes followed by incubation for 24 h in a serum-free, growth factor-free, antibiotic-free 1: 1 (v/v) mixture of Dulbecco's modified Eagle's media and Ham's F-12 (DMEM-F-12; ICN) supplemented with 5 μg ml⁻¹ human transferrin (basic media). Incubations were then performed in DMEM-F-12 supplemented with 10% control or post-nephrectomy serum for 24 h.

Cytological examination of cytocentrifuge preparations of cultured PTCs and CFs from all donors failed to reveal any evidence of cellular atypia. The morphological, biochemical and functional characteristics of these cells have been previously studied in this laboratory and found to reproducibly exhibit the features of PTCs and CFs in vivo (Johnson et al. 1997a,c).

Experimental protocol

The renotropic modulation of PTC growth and apical NHE activity was initially confirmed by comparing the effects of pre- and post-operative sera collected at various time points from nephrectomized and control patients on PTC thymidine incorporation, cellular protein content and apical NHE activity. Cell specificity of the renotropic actions were examined at the same time by assessing the effects of post-nephrectomy and control sera on CF growth. Further experiments were designed to determine whether nephrectomy was associated with specific alterations in the major circulating IGFBPs (i.e. IGFBP-1, -2 and -3) and whether exposure to nephrectomized sera altered PTC-associated IGFBPs and IGF-I binding to PTCs.

Growth parameter measurements

Tritiated thymidine incorporation, an index of DNA synthesis (Fine, 1986), was measured as described previously (Johnson et al. 1997b). Estimation of cellular protein content, an indicator of cell size or hypertrophy (Fine, 1986), was performed on aliquots after cell disruption by a Boy 1000 ultrasound (Rudolf Grauer, Switzerland). Protein concentration was measured by the Bio-Rad protein assay kit II (Bio-Rad, Hercules, CA, USA), using bovine serum albumin as the standard, and corrected for cell number.

Measurement of NHE activity

Apical NHE activity was examined directly in PTCs grown on permeable filters, as described previously (Johnson et al. 1997a), following 24 h incubation in nephrectomized or control sera. Cells were initially acidified by a 20 mmol l⁻¹ NH₄Cl prepulse. Apical NHE activity was then measured as the ethylisopropylamiloride (EIPA)-sensitive component of ²²Na⁺ uptake during the first minute following the apical addition of uptake buffer containing (mmol l⁻¹): 135 NaCl, 4 KCl, 1.2 CaCl₂, 0.8 MgCl₂, 5 D-glucose, 28.3 Hepes, 17.7 Tris base, with 100 μmol l⁻¹ ouabain (Sigma) and 0.2 μCi carrier-free ²²Na⁺ (1 mCi ml⁻¹, specific activity approximately 50 Ci mmol⁻¹; NEN Research Products, Du Pont, Wilmington, DE, USA). Basolateral NHE activity was inhibited by the simultaneous basolateral addition of Na⁺-free buffer, which was identical to uptake buffer, except for replacement of NaCl by an equiosmolar amount of choline chloride. Apical NHE results were adjusted for cellular protein content.

Biochemical measurements

Sodium, potassium, chloride, bicarbonate, urea and creatinine concentrations were determined by standard automated methods (BM/Hitachi 747 autoanalyser; Boehringer-Mannheim) in plasma samples collected concurrently with sera. Plasma osmolality was measured by vapour pressure elevation (model 5100C osmometer; Wescor, Logan, UT, USA).

In view of the likelihood of changes in peripheral renin/angiotensin/aldosterone axis activity post-operatively and of their potentially confounding influence on renal growth (Wolf, 1994), plasma renin activity and aldosterone were measured by commercially available radioimmunoassays (RIA) (NEN Dupont angiotensin I ¹²⁵I radioimmunoassay kit, Amrad Pharmacia Biotech, Uppsala, Sweden; Aldosterone RIA, Sorin, Saluggia, Vercelli, Italy), as described previously (Pollock et al. 1996). Intra-assay coefficients of variation were 8.2 and 2.6%, respectively. Corresponding interassay coefficients of variation were 13.1 and 12.5%.

Total IGF-I concentration in each serum sample was measured following acid-ethanol extraction of IGFBPs by a previously described RIA (Baxter, Brown & Turtle, 1982). Specific, previously published RIAs were also performed to measure serum levels of IGFBP-1 (Baxter, Holman, Corbould, Stranks, Ho & Braund, 1995), IGFBP-2 (Baxter et al. 1995) and IGFBP-3 (Baxter & Martin, 1986). For all IGFBP and IGF-I RIAs, the intra- and interassay coefficients of variation were less than 5 and 10%, respectively.

IGF-I binding to monolayers

In a subset (n = 4) of PTCs exposed to control and nephrectomized sera, IGF-I binding to PTC membranes was quantified according to a previously described technique (Johnson et al. 1997c). Using the chloramine-T method, 5 μg of recombinant human IGF-I was iodinated with 1 mCi ¹²⁵I. The specific activity of ¹²⁵I-IGF-I prepared in this fashion was approximately 200 μCi μg⁻¹. After three washes with ice-cold phosphate-buffered saline (PBS), the basolateral aspects of PTC monolayers, which express the bulk of IGF-I binding sites (Johnson et al. 1997c), were incubated in 300 μl basic media, containing ¹²⁵I-IGF-I (200 000 counts min⁻¹) in the presence or absence of various concentrations (10⁻¹¹-10⁻⁶ mol l⁻¹) of unlabelled IGF-I. Basic media was added to the contralateral apical compartment. All incubations were at 4°C for 2 h. Media were then collected and further radioligand binding to PTC monolayers was stopped by three washes with ice-cold PBS. Cells were solubilized in 200 μl 0.2 mol l⁻¹ NaOH and were subsequently neutralized with an equal volume of 0.2 mol l⁻¹ HCl. Aliquots were submitted for protein determinations and gamma-counting. Non-specific binding, defined as the amount of bound ¹²⁵I-IGF-I in the presence of an excess of unlabelled IGF-I (1 μmol l⁻¹), was subtracted from total binding to determine specific IGF-I binding. Results were adjusted for cellular protein content. ¹²⁵I-IGF-I degradation was determined by precipitation of 50 μl media with an equal volume of 20% TCA (Johnson et al. 1997c). Transepithelial leakage of ¹²⁵I-IGF-I was also measured during each study and was consistently less than 2%.

In view of the fact that IGF-I binds to both cell-associated IGF-I binding proteins (IGFBPs) and IGF-I receptors on PTC surfaces (Bach, Cox, Mendelsohn, Herington, Werther & Jerums, 1992), equilibrium binding studies were also performed using radiolabelled and unlabelled LR³-IGF-I (long arginine³-IGF-I; GroPep, Adelaide, Australia), an IGF-I analogue which binds to the IGF-I receptor, but not to IGFBPs (Ballard, Wallace, Francis, Tomas & Read, 1996).

Statistical analysis

Results are expressed as means ±s.e.m. For the purposes of analysis, each experimental result for the growth studies was expressed as a change from the pre-operative serum value, which was regarded as 100%, and analysed independently. Statistical comparisons between groups at identical time points were made by Student's unpaired t test, whilst comparisons of measurements for sera collected at different time points within each group were made by repeated measures analysis of variance (RM-ANOVA). Differences in gender composition between the two groups were assessed by Fisher's exact test. Analyses were performed using the software package, Statview version 4.5 (Abacus Concepts Inc., Berkeley, CA, USA). P values less than 0.05 were considered significant.

RESULTS

Patient demographics

Patients in the nephrectomy and control groups were well matched for age (64.1 ± 4.0 vs. 61.9 ± 5.8 years, P = 0.75), gender (male: female ratios 6:4 vs. 5:2, P = 0.98) and body mass index (25.1 ± 0.6 vs. 25.0 ± 0.6 kg m⁻², P = 0.96). All females were post-menopausal, except for one patient in the nephrectomy group, and were not receiving hormone replacement therapy. Apart from their conditions which mandated elective surgical intervention, patients were otherwise healthy. The minimum duration of hospitalization was 7 days for both groups.

Effect of sera on PTC and CF growth

Compared with pre-nephrectomy sera, day 3 and day 7 post-nephrectomy sera induced significant increases in both PTC protein contents and thymidine incorporation rates (Figs 1 and 2). In contrast, post-operative sera from control patients resulted in no change in PTC growth compared with pre-operative sera. The absolute values for cellular protein content and thymidine incorporation were not significantly different between PTCs exposed to pre-operative control sera (624 ± 67 pg cell⁻¹ and 2322 ± 307 d.p.m. (1000 cells)⁻¹, respectively) and pre-nephrectomy sera (718 ± 54 pg cell⁻¹, P = 0.28 and 1953 ± 140 d.p.m. (1000 cells)⁻¹, P = 0.26, respectively). Cell numbers were not significantly different between or within groups at any time point (data not shown).

Figure 1. Effect of control and nephrectomized sera on human PTC DNA synthesis.

Thymidine incorporation was measured in confluent, quiescent PTCs incubated for 24 h with pre-operative (day 0) and days 1, 3 or 7 post-operative sera from control (□) and nephrectomized (▪) patients. Results are expressed as a percentage of pre-operative values. The post-operative increase in thymidine incorporation in the nephrectomized group was significant by RM-ANOVA. *P < 0.05versus control.

Figure 2. Effect of control and nephrectomized sera on human PTC hypertrophy.

Cellular protein content was measured in confluent, quiescent PTCs incubated for 24 h with pre-operative (day 0) and days 1, 3 or 7 post-operative sera from control (□) and nephrectomized (▪) patients. Results are expressed as a percentage of pre-operative values. The post-operative increase in protein content in the nephrectomized group was significant by RM-ANOVA. *P < 0.05versus control.

Similarly, CF numbers were not significantly affected by peri-operative sera from either control or nephrectomized patients. However, CF thymidine incorporation was progressively inhibited by both post-operative control and nephrectomized sera, reaching nadir values at day 7 of 81.0 ± 3.7 and 81.5 ± 6.8% of pre-operative controls, respectively.

Effect of sera on PTC apical NHE activity

Baseline values for apical NHE activities were not different between the two groups (Fig. 3). However, there was a significant increase in NHE activity in PTCs exposed to day 1 post-nephrectomy sera. This was followed by a return to baseline values at days 3 and 7, which coincided with the time points at which PTC cellular protein contents were appreciably increased. Conversely, post-operative sera from control patients exerted no stimulatory effect on PTC apical NHE compared with pre-operative sera.

Figure 3. Effect of control and nephrectomized sera on human PTC apical NHE activity.

Apical NHE activity, measured as the component of apical ²²Na⁺ uptake inhibited by 10 μmol l⁻¹ EIPA, was measured in confluent, quiescent PTCs incubated for 24 h with pre-operative (day 0) and days 1, 3 or 7 post-operative sera from control (□) and nephrectomized (▪) patients. Results were adjusted for cellular protein content. The initial post-operative increase in NHE activity in the nephrectomized group was significantly different from pre-nephrectomy values. *P < 0.05versus control.

Biochemical measurements

Pre- and post-operative plasma biochemistries were similar between the two groups (Table 1). No differences reached statistical significance, apart from the plasma potassium on day 3. Pre- and post-operative urea and creatinine concentrations tended to be slightly higher in nephrectomized patients, but did not change significantly over time relative to control patients. Plasma sodium and osmolality fell significantly and to an equivalent extent in both groups post-operatively. These findings were mirrored by significant decreases in both plasma renin activities and aldosterone levels in control and nephrectomized patients on the first and third post-operative days, followed by recovery on day 7.

Table 1.

Characteristics of plasma obtained from control (C) and nephrectomized (N) patients at the indicated peri-operative time points

Parameter	Day 0		Day 1		Day 3		Day 7
	C	N	C	N	C	N	C	N
Na+	140 ± 1	139 ± 1	137 ± 1	138 ± 1	136 ± 1	136 ± 0	137 ± 1	136 ± 1
K+	4.2 ± 0.1	4.3 ± 0.1	4.3 ± 0.2	4.3 ± 0.2	4.0 ± 0.0	4.4 ± 0.1*	4.1 ± 0.1	4.3 ± 0.2
Cl−	102 ± 1	103 ± 2	99 ± 3	104 ± 1	101 ± 2	101 ± 1	100 ± 2	100 ± 1
HCO3−	27 ± 1	26 ± 1	27 ± 1	25 ± 1	28 ± 1	27 ± 1	26 ± 1	27 ± 0
Urea	4.8 ± 0.7	7.2 ± 1.1	4.4 ± 0.7	6.1 ± 0.7	3.9 ± 0.6	6.2 ± 0.9	4.2 ± 0.6	6.5 ± 1.1
Creatinine	0.08 ± 0.01	0.10 ± 0.02	0.08 ± 0.01	0.12 ± 0.02	0.08 ± 0.01	0.12 ± 0.02	0.08 ± 0.01	0.12 ± 0.02
Osmolality	305 ± 8	298 ± 2	290 ± 3	294 ± 2	289 ± 2	292 ± 1	290 ± 2	292 ± 1
PRA	1.58 ± 0.46	1.95 ± 0.38	1.40 ± 0.38	1.00 ± 0.36	0.84 ± 0.35	0.56 ± 0.18	1.62 ± 0.73	1.90 ± 0.37
Aldosterone	400 ± 94	272 ± 43	179 ± 45	227 ± 56	213 ± 85	56 ± 13	135 ± 86	354 ± 93
IGF-I	125 ± 52	100 ± 16	117 ± 42	100 ± 13	99 ± 41	79 ± 7	98 ± 37	113 ± 20
IGFBP-1	29 ± 7	69 ± 18	106 ± 63	108 ± 41	98 ± 25	87 ± 27	62 ± 25	59 ± 20
IGFBP-2†	1180 ± 410	1230 ± 263	1940 ± 518	2820 ± 571	1930 ± 519	2730 ± 462	2800 ± 1010	3650 ± 803
IGFBP-3†	3580 ± 305	3360 ± 217	3060 ± 349	2840 ± 259	2860 ± 323	2630 ± 234	2670 ± 341	2602 ± 347

PRA, plasma renin activity. Units are mmol l⁻¹, except for osmolality (mosmol kg⁻¹), PRA (ng ml⁻¹ h⁻¹), aldosterone (fmol (ml protein)⁻¹), IGF-I (ng ml⁻¹) and IGFBPs (ng ml⁻¹).

^*P < 0.05versus control.

^†Significant post-operative changes in concentrations in both groups by RM-ANOVA.

Serum IGF-I and IGFBP-1 levels were not significantly different over time or between groups (Table 1). However, IGFBP-3 levels decreased progressively (P < 0.0001) and to an equivalent extent from comparable baseline values in both post-operative control and post-nephrectomy patients. In contrast, IGFBP-2 levels increased progressively (P < 0.001) and to an equivalent extent from comparable baseline values in both groups.

Effect of sera on IGF-I and LR³-IGF-I binding to PTC membranes

¹²⁵I-IGF-I binding to PTC membranes was comparable following exposure to pre-nephrectomy and pre-operative control sera (142 ± 4 versus 144 ± 7 fmol (mg protein)⁻¹, n.s.) (Fig. 4). Post-operatively, ¹²⁵I-IGF-I binding to PTC membranes increased significantly only in cells exposed to post-nephrectomy sera. Subsequent amounts of binding in the post-nephrectomy group at days 3 and 7, following correction for their increased cellular protein contents at these time points, were not significantly different from controls or from pre-nephrectomy levels. The changes in specific binding in the nephrectomy group could not be explained by differential rates of ¹²⁵I-IGF-I degradation (day 0, 5.3 ± 1.8%; day 1, 4.6 ± 1.0%; day 3, 4.5 ± 1.9%; day 7, 4.0 ± 1.6%; n.s.).

Figure 4. Effect of peri-operative control and nephrectomized sera on specific binding of ¹²⁵I-IGF-I to PTCs.

Confluent, quiescent PTC monolayers were incubated for 24 h in 10% peri-operative sera from control (□, n = 4) and nephrectomized (▪, n = 4) patients. After washing, ¹²⁵I-IGF-I (200 000 counts min⁻¹) was added for 2 h at 4 °C in the presence or absence of an excess (1 μmol l⁻¹) of unlabelled IGF-I. The amount of specifically bound ¹²⁵I-IGF-I (generally 8-10% of total tracer counts) was adjusted for cellular protein content. *P < 0.05versus control.

In contrast to the finding of increased ¹²⁵I-IGF-I binding to PTC membranes following exposure to post-nephrectomy sera, specific binding of ¹²⁵I-LR³-IGF-I to PTC surfaces was not significantly altered by pre-incubation with post-nephrectomy sera (day 0, 29 ± 2 fmol (mg protein)⁻¹; day 1, 28 ± 2 fmol (mg protein)⁻¹; day 3, 28 ± 2 fmol (mg protein)⁻¹; day 7, 26 ± 2 fmol (mg protein)⁻¹; n.s.). Similarly, ¹²⁵I-LR³-IGF-I degradation rates were unaffected (8.0 ± 2.3, 6.2 ± 4.6, 8.2 ± 4.0 and 8.6 ± 3.7%, respectively; n.s.).

DISCUSSION

The present study clearly demonstrates the presence of a humoral factor(s) in human serum induced by unilateral nephrectomy, which specifically stimulates hypertrophy and, to a lesser extent, hyperplasia of human proximal tubule cells (PTCs). This growth stimulation is preceded by a transient increase in apical sodium-hydrogen exchange (NHE) activity and specific binding of IGF-I to PTCs. Although post-nephrectomy sera exhibited profound changes in the profile of circulating IGFBPs compared with pre-nephrectomy sera, a renotropic role for circulating IGFBPs in humans could not be substantiated, since comparable changes were found in IGFBP profiles of post-operative sera drawn from control patients.

In the current study, the maximum increase in PTC protein content induced by post-nephrectomy sera (approximately 30% above pre-nephrectomy values) was similar to that reported in vivo for compensatory renal growth following unilateral nephrectomy in both animals (Hayslett, 1983; Polychronakos et al. 1985; Flyvbjerg et al. 1989) and humans (Hayslett, 1983). It appears likely that this growth-stimulatory effect relates specifically to the removal of normal functional tissue rather than renal neoplasms, because sera collected from a subgroup of patients undergoing nephrectomy for benign lesions tended to augment PTC protein contents and thymidine incorporation rates to a greater extent than sera obtained from patients with renal malignancies. Similar findings have also been described by Yamamoto et al. (1983).

The renotropic activity of human post-nephrectomy sera was specific for PTCs, since the growth of renal cortical fibroblasts was inhibited, rather than stimulated. Although this phenomenon has not been previously described, the growth of primary rabbit skin fibroblasts has been found to be inhibited by approximately 10% in the presence of nephrectomized rabbit serum (Yamamoto et al. 1983). Moreover, renal fibroblast contamination of PTC cultures has been suggested to result in attenuation of the net growth-stimulatory effects of post-nephrectomy serum (Yamamoto et al. 1983; Yun et al. 1988). In vivo studies by other investigators (Williams, 1961; Lowenstein & Stern, 1963) further indicate that renotropin stimulates growth of proximal tubule epithelial cells, but not of renal fibroblasts. Physiologically, such a mechanism may serve to protect against the development of tubulointerstitial fibrosis accompanying adaptive increases in tubular growth and function.

In keeping with in vitro and in vivo studies of compensatory renal growth in animals (Fine, 1986; Pollock & Field, 1993; Pollock et al. 1996), stimulation of hypertrophy by post-nephrectomy serum was antedated by an augmentation of apical NHE activity. Serum from nephrectomized rats has similarly been shown to stimulate proximal tubule Na⁺ reabsorption in rats in vivo (Pollock et al. 1996) and to enhance Na⁺ transport in LLC-PK₁ cells in vitro (Esbrit et al. 1991). Augmentation of NHE activity has been suggested to promote tubular growth directly by raising intracellular Na⁺ or pH, or indirectly via secondary increases in K⁺ and Ca²⁺ (Grinstein, Rotin & Mason, 1989; Stanton & Kaissling, 1989). Our results suggest that a circulating factor may promote an increase in apical NHE activity prior to cell growth. However, several investigators have reported a dissociation of PTC hypertrophy and NHE activity under certain experimental conditions (Grinstein et al. 1989; Mackovic Basic, Fan & Kurtz, 1992).

The increase in NHE activity prior to cellular hypertrophy was associated with a parallel increase in IGF-I binding to human PTCs. Since IGF-I stimulates PTC growth and apical NHE activity both in vitro and in vivo (Zumkeller & Schofield, 1992; Hammerman & Miller, 1993; Nobes et al. 1995; Johnson et al. 1997b), it is possible that an increase in binding of, and response to, ambient IGF-I contributed to the stimulation of apical NHE and initiation of PTC growth by post-nephrectomy sera in the present study. In keeping with this finding, Polycronakos et al. (1985) observed a significant 59% increase in binding of radiolabelled IGF-I to microsomal membranes in uninephrectomized rats prior to the onset of kidney growth. Furthermore, Hise et al. (1993) showed that once compensatory renal growth is established in rats, the increase in proximal tubule IGF-I binding is matched by a commensurate increase in cellular protein content, such that the number of IGF-I binding sites adjusted for cellular protein was not significantly different from sham-operated controls. Our results would suggest that a similar process occurs in humans.

The competition characteristics of radiolabelled LR³-IGF-I versus IGF-I binding to PTCs following incubation with post-nephrectomy sera imply that the initial increase in specific binding of IGF-I reflects an increase in either the number or affinity of cell-associated IGFBPs, rather than type I IGF receptors, on PTC surfaces. We have previously demonstrated cell-associated IGFBP-3 in human PTCs by immunoblot analysis and found that cell association of this binding protein potentiates the action of IGF-I on PTCs, possibly through facilitation of interaction with its receptor (Johnson et al. 1997c). This mechanism has also been described by other authors (De Mellow & Baxter, 1988; Conover, 1992) and has been implicated in diabetic proximal tubular hypertrophy (Bach et al. 1992). The only prior study of the expression or role of IGFBPs in contralateral kidneys following unilateral nephrectomy was by Hise et al. (1993), who demonstrated that established compensatory proximal tubular hypertrophy in rats was indeed associated with an increase in cell-associated IGFBP in proportion to membrane protein content and surface area. Unlike human PTCs, however, the cell-associated IGFBP was identified by immunostaining as IGFBP-5.

Circulating IGF-I and IGFBPs were also evaluated in post-nephrectomy human sera in the present study to determine whether changes in these components contributed to renotropic stimulation of PTC growth and transport. Like most other studies (Stiles et al. 1985; Fagin & Melmed, 1987; Lajara et al. 1989; Hammerman & Miller, 1993; Gronboek et al. 1997), we observed that total serum IGF-I was unaffected by nephrectomy. However, since serum IGFBPs play a crucial role in determining IGF-I bioavailability to tissues (Jones & Clemmons, 1995; Bereket et al. 1996) and since the relative amounts of each of the different soluble IGFBPs may further be important in targeting the organ and cell specificity of IGF-I action (Minuto, Barreca, Del Monte & Giordano, 1991; Jones & Clemmons, 1995), we reasoned that nephrectomy may facilitate compensatory renal growth by inducing a distinctive pattern of changes in serum IGFBPs, which in turn promoted specific repartitioning of IGF-I from the circulation to human proximal tubules. Although nephrectomy was in fact followed by a progressive fall in serum IGFBP-3, an increase in IGFBP-2 and a non-significant biphasic rise and then fall in IGFBP-1, almost identical changes were seen in the control group, thereby indicating that alterations in the serum IGF-I/IGFBP axis were unlikely to account for the renotropic activity of nephrectomized sera. This contrasts with the only prior study of serum IGFBPs following unilateral nephrectomy (Gronboek et al. 1997), which found no significant changes in serum IGFBP-1 or IGFBP-3 during the week following graded renal ablation in rats. The apparent disparity in results may reflect differences in peri-operative factors or species-specific responses, since post-operative stress following cardiac (Holly, Claffey, Cwyfan Hughes, Frost & Yateman, 1993) or abdominal surgery (Cotterill et al. 1996) in man has been reported to consistently suppress serum IGFBP-3 to an extent equivalent to that described in the present study. This suppression has been attributed to non-specific induction of IGFBP proteases (Holly et al. 1993).

To our knowledge, there have only been four previous studies of renotropic activity in human sera following nephrectomy (Yamada et al. 1983; Yamamoto et al. 1983; Memon, Gongwei, Ahmed, Gilbert, Regan & Preuss, 1993; Garcia Ocana, Ortega, Gonzales-Garcia, Garcia-Canton & Esbrit, 1993). Similar to the present investigation, these studies observed maximal stimulation of human PTC thymidine incorporation by post-nephrectomy sera at approximately day 7, although the degree of peak stimulation varied considerably between 30% (Memon et al. 1993) and 600% (Yamamoto et al. 1983). The latter differences may be explained by patient heterogeneity and by differences in cell density, and therefore in cell-contact inhibition of proliferation, at the time of serum stimulation (Yamamoto et al. 1983).

The major drawbacks of each of these four studies included the use of human PTC cultures established from only one (Yamamoto et al. 1983) or two kidneys (Yamada et al. 1983; Memon et al. 1993), lack of sufficient characterization of cultures to exclude contamination by adventitious cells (e.g. fibroblasts or distal tubular epithelial cells) and the failure to include post-operative sera from appropriate controls. The latter point is particularly important since some investigators have demonstrated a degree of renotropic activity in the sera of sham-operated animals (Logan & Benson, 1992; Garcia Ocana & Esbrit, 1994). Moreover, the sole end-point for quantitation of renotropic activity in these studies was thymidine incorporation, which seems inappropriate given that the predominant component of compensatory proximal tubular growth involves hypertrophy, rather than hyperplasia (Hayslett, 1983; Fine, 1986).

Although the activity of the peripheral renin/angiotensin/ aldosterone axis is increased following nephrectomy in rats (Pollock et al. 1996) and has been linked to proximal tubule growth in animal cell lines (Wolf, 1994), in our study, both plasma renin activities and aldosterone levels were found to be suppressed in humans in both the nephrectomized and control groups. The suppression most probably resulted from post-operative volume expansion and argues against a significant role of the peripheral renin/angiotensin/ aldosterone axis in mediating compensatory renal growth.

In conclusion, the present study demonstrates that human patients undergoing nephrectomy elaborate a serum factor(s), which specifically stimulates human proximal tubule apical NHE activity and subsequent hypertrophy and DNA synthesis in vitro. This effect may be partly attributable to an early increase in binding of circulating IGF-I to cell-associated IGFBPs on PTC surfaces, although no change in the circulating IGFBP profile may be demonstrated. The observed alterations in growth appear specific to the proximal tubule cell as no stimulation of cortical fibroblast growth occurs.