sign in sign up
<<
Home , Erythropoietin, Growth factor

J Am Soc Nephrol 10: 315–322, 1999 -Like I Plays a Role in Regulating in Patients with End-Stage Renal Disease and Erythrocytosis

LEE-YUNG SHIH,* JENG-YI HUANG,† and CHING-TAI LEE* *Division of - and †Division of , Department of Internal Medicine, Chang Gung University, Chang Gung Memorial Hospital, Taipei, Taiwan.

Abstract. Erythroid progenitor growth, the dent erythroid colonies were found. Serum EPO levels were that regulate erythropoiesis, and the effect of patient’s serum constantly normal in one patient and elevated in three patients on the growth of normal erythroid progenitors were assessed in with ACDK; for non-ACDK patients, EPO levels were normal eight patients with end-stage renal disease (ESRD) and eryth- or low in two patients and persistently increased in one, but rocytosis. All patients were male and had been on maintenance fluctuated in the remaining one on serial assays. There was no , they had a Ͼ50% and/or a red blood correlation between serum EPO levels and hematocrit values. count Ͼ6 ϫ 1012/L and an arterial oxygen saturation Ͼ95%. The serum IGF-I levels in patients with ESRD and erythrocy- Four had acquired cystic disease of the (ACDK), and tosis were significantly increased compared with normal sub- four other non-ACDK patients did not have known causes of jects or ESRD patients with . We found an inverse secondary erythrocytosis after appropriate investigations and correlation between serum EPO and IGF-I levels. Sera from long-term follow-up. The methylcellulose culture technique patients with ESRD and erythrocytosis exhibited a stimulating was used to assay the erythroid progenitor (BFU-E/CFU-E) effect on normal marrow CFU-E growth. The stimulating growth. Serum (EPO) and insulin-like growth effect of sera from patients who had a normal serum EPO level factor I (IGF-I) levels were measured by RIA. Paired experi- and an elevated IGF-I level could be partially blocked by ments were performed to determine the effects of 10% sera anti-IGF-I. The present study suggests that IGF-I plays an from ESRD patients and control subjects on normal marrow important role in the regulation of erythropoiesis in patients CFU-E growth. The numbers of EPO-dependent BFU-E in with ESRD and erythrocytosis who did not have an increased marrow and/or blood of patients with ESRD and erythrocytosis EPO production. were higher than those of normal controls. No EPO-indepen-

Erythrocytosis may complicate the disease course of renal erythrocytosis and ESRD on maintenance dialysis: four had tumor (1), cysts (2), polycystic kidney (3), hydronephrosis (4), ACDK and the remaining four did not fit into renal artery stenosis (5), and postrenal transplant (6,7). In most vera or any known mechanism of secondary erythrocytosis. of these patients, the renal function was normal or only mildly We assessed the in vitro erythroid progenitor growth, serum impaired. On the contrary, anemia of varying degrees is in- erythropoietin (EPO), and insulin-like growth factor I (IGF-I) variably present in patients with chronic renal failure except for levels, and the effects of sera from uremic patients on the some patients with polycystic kidney disease (8,9). The pres- growth of normal erythroid progenitors. Our results suggest ence of erythrocytosis in the setting of end-stage renal disease that IGF-I is an important regulator in some patients with (ESRD) is an extremely rare phenomenon. There have been ESRD and erythrocytosis. To the best of our knowledge, IGF- rare reports that described only one or two patients with ESRD I-mediated erythrocytosis in patients with ESRD has not been and erythrocytosis of a variety of etiologies including acquired described previously. cystic disease of the kidney (ACDK) (10), secondary erythro- cytosis due to (11), polycystic kidney disease with Materials and Methods renal artery thrombosis (12), diabetes mellitus (13), and poly- Patient Population cythemia vera (14–16). We encountered eight patients with From 1982 to 1996, 13 patients undergoing hemodialysis were found to be associated with erythrocytosis at Chang Gung Memorial Received September 15, 1997. Accepted August 14, 1998. Hospital, where the Hemodialysis Center serves 850 patients for Correspondence to Dr. Lee-Yung Shih, Divison of Hematology-Oncology, regular dialysis. Of them, five with clinical and laboratory features Department of Internal Medicine, Chang Gung Memorial Hospital, 199 Tung that fulfilled the diagnostic criteria of were ex- Hwa North Road, Taipei, Taiwan. Phone: ϩ886 3 3281200, extension 2524; cluded (17); the remaining eight patients constituted the basis of this Fax: ϩ886 3 3286697; E-mail: [email protected] study (patients 5 and 6 were reported previously) (18). All of our 1046-6673/1002-0315$03.00/0 patients were male and had a hematocrit Ͼ50% and/or a 12 Journal of the American Society of Nephrology count Ͼ6 ϫ 10 /L on two or more consecutive visits, an arterial Copyright © 1999 by the American Society of Nephrology oxygen saturation Ͼ95% on at least two measurements, and a blood 316 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 315–322, 1999 urea nitrogen Ͼ100 mg/dl or serum creatinine level Ͼ10 mg/dl. All stored at Ϫ70°C before testing. In the paired experiments, the effect patients had function tests carried out each month during dialysis of 10% sera from both patients and control subjects on normal therapy. None had pulmonary disease and all denied prior recombi- erythroid progenitor (CFU-E) growth was evaluated using erythroid nant human EPO therapy. ␣-fetoprotein level, P50 measurement, and progenitor assay in the presence of 1 U EPO. The amount of fetal calf liver and renal sonograms were performed in all patients. Carboxy- serum was reduced to 20% to maintain the total serum concentration level was determined in the smokers. of 30%. The experiment was repeated on four different normal target marrows for each serum sample. The effect of serum from uremic Erythroid Progenitor Cell Assay patients on normal CFU-E colony growth was expressed as a mean Mononuclear cell fraction of heparinized marrow (BM) and percentage of that with the addition of normal serum. In addition, sera peripheral blood (PB) cells were isolated by Ficoll-Hypaque (1.077 from 11 undialyzed ESRD patients with anemia immediately before g/ml) density gradient centrifugation, and the nonadherent (NA) frac- the onset of regular hemodialysis were also studied for comparison, tion was separated by adherence technique (19). The NA cells were and sera from another 13 patients undergoing maintenance hemo- depleted of E-rosette-forming cells by second Ficoll-Hypaque centrif- dialysis for at least 6 mo were assayed before and after one hemodi- ugation of a mixture of NA cells and sheep red blood cells. T-depleted alysis session. cells were obtained at the interface and cultured in dishes using Iscove’s technique with modification (20,21). Briefly, 5 ϫ 104 NA Anti-IGF-I Neutralizing Experiment T-depleted cells were plated in 0.9% methylcellulose, supplemented Polyclonal antihuman IGF-I (Pepro Tech, London, United King- with Iscove’s modified Dulbecco’s medium, 5 ϫ 10Ϫ5 M 2-mercap- dom) was used in the neutralizing experiment. Sera were preincubated toethanol, 1% bovine serum albumin, 30% fetal calf serum, with or with 10 ␮g/ml anti-IGF-I overnight before addition to the cultures for without addition of 1 U EPO (Connaught Laboratory, Willowdale, in vitro erythroid progenitor assay as described above. The neutraliz- Ontario, Canada). Dishes were incubated at 37°C in a humidified ing effect was expressed as the percentage of CFU-E number with the atmosphere of 5% CO2 in air. Erythroid colonies were scored by their addition of anti-IGF-I to that without anti-IGF-I in the presence of red color on an inverted microscope on day 7 for CFU-E (colony uremic serum in normal marrow culture. forming unit-erythroid) and day 14 for BFU-E (burst forming unit- erythroid). Well hemoglobinized colonies visible at 14 d of incubation Statistical Analyses in the absence of added EPO were counted as EPO-independent Because the sample size was small, comparison of variables erythroid colonies. Thirty normal BM samples, and 20 PB from (BFU-E numbers, serum EPO, and IGF-I levels) between two groups healthy subjects served as control subjects. Four patients with renal was made by the nonparametric Wilcoxon sum test. Comparison failure and anemia served as an additional control group. Informed of the effects of sera from uremic patients and healthy subjects as well consent was obtained from each patient and control subject; investi- as the effects before and after dialysis in the paired experiments was gations were approved by the Human Research Committee at Chang made by Wilcoxon signed rank test. Spearman’s correlation coeffi- Gung Memorial Hospital. cient was used to assess the linear relationship and its variability between two variables. Determination of Serum EPO Concentration, Serum IGF-I Level, and Serum Intact Parathyroid Results Blood samples were taken in the morning at initial evaluation and The patient characteristics and laboratory findings at the Ϫ during the subsequent course. Serum was freshly frozen at 70°C time of investigation are shown in Tables 1 and 2. The causes until assay. Serum EPO levels were measured by RIA using a com- of renal failure included: chronic glomerulonephritis (n ϭ 2), mercially available (Diagnostic Systems Laboratories, Webster, ϭ TX) with rabbit antihuman EPO purified from and calibrated diabetes mellitus (n 1), diabetes mellitus with ϭ ϭ according to the World Health Organization 2nd International Refer- (n 1), and rapidly progressive glomerulonephritis (n 1). ence preparation. The intra-assay coefficient of variation (n ϭ 10) for The remaining three patients initially presented with ESRD EPO levels was 9.8%, and the interassay coefficient of variation (n ϭ with bilateral small and contracted kidneys, for which the 10) was 10.5%. Serum IGF-I levels were also determined by RIA with causes of renal failure could not be defined. None had spleno- a commercially available kit (Nichols Institute Diagnostics, San Juan megaly. All had arterial oxygen saturation Ͼ95% on repeated Capistrano, CA), using acid-ethanol extraction method to dissociate tests, and none had an elevated carboxyhemoglobin level in IGF-I from IGF-binding . The intra-assay coefficients of smokers. The observed counts ranged from 0.8 to ϭ ␮ ␮ variation (n 12) were 7.2% at 40 g/L IGF-I and 4.2% at 157 g/L 2.8%, but the absolute reticulocyte counts were increased in the ϭ IGF-I, and the respective interassay coefficients of variation (n 10) presence of erythrocytosis. Lactate dehydrogenase and hapto- were 10.1 and 6.4%. Serum intact (iPTH) was globin levels were all within normal limits. The leukocyte measured using a commercial RIA kit (Nichols Institute Diagnostics), with a normal range of 10 to 65 pg/ml. Duplicate tests were performed alkaline phosphatase scores were normal in five patients, low for each sample. Sera from healthy adults served as controls. We also in one, and slightly elevated in two. All patients had normal included an additional control group of ESRD patients with anemia, liver function tests and an ␣-fetoprotein level below 5 ng/ml. who had not yet received recombinant human EPO therapy. Liver sonography revealed a normal liver without tumor or cyst in all patients. Absence of marrow iron stores was found in Effects of Sera from Uremic Patients on Normal patients 3 and 5, who also had low serum ferritin levels. Erythropoiesis Renal sonography showed that four patients had ACDK, Serum was obtained from each patient immediately before regular with bilateral multiple renal cysts of various sizes in patients 1, hemodialysis. A pool of sera from 10 healthy volunteers served as 2, and 3, and multiple cysts in the graft kidney in patient 4. The control serum. Serum was heat-inactivated at 56°C for 30 min and sizes of the kidneys with cystic lesions were small in patients J Am Soc Nephrol 10: 315–322, 1999 IGF-I in Uremia and Erythrocytosis 317

Table 1. Patient characteristics and features of renal disease at the time of study

Duration of Patient Age/Gender Cause of Renal Dialysis Findings on Renal Sonogram Failure (months)

1 46/M Chronic 82 Bilateral small kidneys (right 8.2 ϫ 4.0 cm, glomerulonephritis left 8.5 ϫ 4.1 cm) with multiple (Ͼ20) renal cysts 2 39/M Unknown 24a Bilateral small and contracted kidneys (right 5.9 ϫ 3.3 cm, left 6.5 ϫ 3.1 cm) with multiple (7) renal cysts 3 32/M Rapidly progressive 36a Bilateral normal-sized kidneys (right 11.9 glomerulonephritis ϫ 6.3 cm, left 11.6 ϫ 6.1 cm) with multiple (Ͼ20) renal cysts 4 41/M Unknown 2 Bilateral small native kidneys (right 5.3 ϫ 2.6 cm, left 5.3 ϫ 3.2 cm) with multiple (6) renal cysts in the enlarged renal graft (13.1 ϫ 6.4 cm) 5 51/M Diabetes mellitus 45 Bilateral small kidneys (right 8.4 ϫ 4.0 cm, left 8.8 ϫ 4.3 cm) 6 50/M Diabetes mellitus, 50 Bilateral small kidneys (right 8.3 ϫ 4.2 cm, hypertension left 8.1 ϫ 3.8 cm), a small stone in right kidney 7 45/M Chronic 35 Bilateral small kidneys (right 9.0 ϫ 4.1 cm, glomerulonephritis left 8.7 ϫ 4.3 cm) 8 43/M Unknown 4 Bilateral small and contracted native kidneys (right 7.0 ϫ 3.3 cm, left 6.6 ϫ 4.0 cm), a 0.8-cm cyst in left kidney; normal-sized renal graft (10.1 ϫ 4.1 cm) with parenchymal disease echo-pattern and no cyst

a Following graft removal due to chronic rejection.

1 and 2, and normal in patient 3; patient 4 had a slightly ocrit levels remained elevated. Renal sonography of patient 8 enlarged renal graft. Patients 2 and 3 had undergone renal revealed small and contracted native kidneys with a single transplant before, but both had graft removal due to rejection. small cyst measuring 0.8 cm in the left kidney that had been They had been on maintenance dialysis 2 and 3 yr, respec- present before transplantation, and no cyst in the renal graft. tively, before the development of erythrocytosis. Patient 4 Patients 5, 6, 7, and 8 did not have leukocytosis or thrombo- received renal transplantation in December 1987, which was cytosis. P50 and B12 levels were all within normal limits, and followed by chronic rejection; multiple renal cysts in the graft no known causes of secondary erythrocytosis were found in the kidney were found 28 mo later, the renal cysts progressively four patients of non-ACDK during a follow-up time ranging enlarged, while both native kidneys remained small and con- from 14 to 56 mo. tracted without cysts in the subsequent sonography follow-up. The hematocrit levels of patient 4 ranged between 50.3 and 59.7% in the following 8 yr. He has been on regular hemodi- Erythroid Progenitor Culture Assays alysis again since May 1995 when the hematocrit level was Table 3 shows the results of in vitro EPO-dependent BFU-E 53.1%. growth. The number of BFU-E was significantly increased in Sonography of the kidneys showed small kidneys without ESRD patients with erythrocytosis than in normal controls tumors, hydronephrosis, or cysts in patients 5, 6, and 7; the (P ϭ 0.021 for BM; P ϭ 0.05 for PB), whereas there was no duration of hemodialysis before the discovery of erythrocytosis statistical difference in the BM BFU-E numbers between was 45, 50, and 35 mo, respectively. Patient 8 underwent renal ESRD patients with erythrocytosis and ESRD with anemia transplant in September 1989. He developed erythrocytosis in (P ϭ 0.5) or between ESRD with anemia and normal controls July 1990, and his graft function began to deteriorate due to (P ϭ 0.092). In the absence of exogenous EPO in cultures, graft rejection in July 1991, which progressed to renal failure none formed EPO-independent erythroid colonies in ESRD in February 1996 when hemodialysis was restarted. His hemat- patients with erythrocytosis or anemia, or in normal controls. 318 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 315–322, 1999

Table 2. Laboratory findings at the time of evaluation in patients with chronic renal failure and erythrocytosisa

AST/ Reticulo- RBC Bone Serum Serum BUN/Cr Hb Hct RBC MCV WBC O2 Sat Marrow b c Patient ALT ϫ 12 cyte (%)/ ϫ 9 ϫ 9 Mass Ferritin iPTH (mg/dl) (g/dl) (%) ( 10 /L) ϫ 9 (fl) ( 10 /L) ( 10 /L) (%) Iron (U/L) ( 10 /L) (ml/kg) Stores (ng/ml) (pg/ml)

1 94/17.8 29/24 17.7 54.7 6.23 2.8/174 87.8 9.5 159 96.2 28.6 Normal 566 79 2 86/13.9 22/32 18.4 55.5 6.60 1.0/66 85.5 5.5 128 97.4 37.7 Normal 158 298 3 79/13.7 19/22 17.4 56.2 7.12 1.4/100 78.9 6.0 190 97.7 48.4 Absent 24.5 473 4 82/13.9 10/17 18.2 58.2 7.20 0.8/58 80.9 8.6 147 95.8 39.2 ND 19.5 1396 5 104/16.0 24/30 16.2 50.6 5.80 1.1/64 87.2 8.1 149 97.0 ND Absent 37 ND 6 83/13.5 33/36 17.6 52.8 5.76 1.2/69 91.6 7.6 172 95.0 ND Mildly 128 ND decreased 7 72/16.4 9/13 17.3 50.9 5.68 0.9/51 89.6 7.2 259 97.9 30.3 Mildly 105 151 decreased 8 101/8.6 32/31 19.3 59.0 6.10 1.5/92 96.7 8.5 153 95.6 46.5 Increased 4289 280

a BUN, blood urea nitrogen; Cr, creatinine; AST, aspartate aminotransferase; ALT, alanine aminotransferase; Hb, hemoglobin; Hct, hematocrit; RBC, red blood cells; MCV, mean corpuscular volume; WBC, white blood cells; O2 Sat, oxygen saturation; iPTH, intact parathyroid hormone; ND, not done. b Normal values for male subjects: mean Ϯ SD ϭ 159.9 Ϯ 84.4 ng/ml; for male uremic patients: 1172 Ϯ 1090 ng/ml (22). c Normal values of serum iPTH: 10 to 65 pg/ml.

Table 3. Erythroid progenitor (BFU-E) growth in patients erythrocytosis (r ϭϪ0.07143, P ϭ 0.87) and uremic anemic with ESRD and erythrocytosisa patients (r ϭϪ0.07762, P ϭ 0.43), respectively. Fresh-frozen sera at initial evaluation were available for No. of BFU-E Patient IGF-I assay in all but patient 5, in whom the IGF-I level was BM PB measured at a later time. Patients with ESRD and erythrocy- tosis had significantly higher IGF-I levels than healthy control 1 318 141 subjects (P ϭ 0.0064) or those with ESRD and anemia (P ϭ 2 654 196 0.0076). The serum IGF-I levels of patients with ESRD and 3 249 105 anemia were lower than those of healthy control subjects, but 4ND33the difference did not reach statistical significance (P ϭ 5 651 187 0.0568). Patients 1, 7, and 8 who had low or normal EPO 6 600 204 concentrations had markedly elevated IGF-I levels. There was 7 1190 167 an inverse correlation between serum EPO and IGF-I levels 88853(r ϭϪ0.8095, P ϭ 0.0149). a Patients: BM (n ϭ 7), BFU-E (mean Ϯ SD) 536 Ϯ 363; PB In the presence of exogenous 1 U EPO, 10% sera from (n ϭ 8), BFU-E (mean Ϯ SD) 136 Ϯ 66. Control subjects, ESRD patients with ESRD and erythrocytosis stimulated the normal with anemia: BM (n ϭ 4), BFU-E (mean Ϯ SD) 362 Ϯ 133. marrow to give rise to CFU-E colonies 114 to 156% (mean Ϯ Normal subjects: BM (n ϭ 30), BFU-E (mean Ϯ SD) 271 Ϯ 160; SD 135 Ϯ 18%) of those formed by 10% normal sera. In ϭ Ϯ Ϯ PB (n 20), BFU-E (mean SD) 88 77. BFU-E, burst contrast, sera from patients with ESRD and anemia were forming unit-erythroid; ESRD, end-stage renal disease; BM, ; PB, peripheral blood; ND, not done. inhibitory to erythropoiesis, CFU-E growth being 26 to 106% (mean Ϯ SD 84 Ϯ 13%) and 63 to 101% (mean Ϯ SD 80 Ϯ 12%) of those supported by normal sera in predialyzed and regularly dialyzed patients, respectively. The differences in the Serum EPO and IGF-I Levels and Effects of Sera from effects of sera on normal CFU-E growth were statistically Patients with ESRD on in Vitro Normal Marrow CFU- significant between ESRD patients with erythrocytosis and E Growth predialyzed uremic anemic patients (P ϭ 0.0003) or between The serum EPO concentrations of patients with ACDK were ESRD patients with erythrocytosis and chronic dialyzed ane- increased in three patients and constantly normal in one, mic patients (P ϭ 0.0002). No statistical difference was found whereas two of the four patients with non-ACDK had normal between predialyzed sera and sera from those on maintenance or low serum EPO levels and one (patient 6) had a persistently dialysis in patients with ESRD and anemia (P ϭ 0.092). Sera elevated value on repeated assays with comparable hematocrit assayed in the same individuals with ESRD and anemia before values during the follow-up course. Patient 5 had fluctuating and after one hemodialysis session showed that the effects on titers, being normal initially but increasing later in the subse- CFU-E growth after dialysis were 82 to 136% (mean Ϯ SD quent measurements. There was no correlation between serum 103 Ϯ 16%) of those before dialysis (P ϭ 0.65). EPO levels and hematocrit values in ESRD patients with The results of neutralizing experiments showed that the J Am Soc Nephrol 10: 315–322, 1999 IGF-I in Uremia and Erythrocytosis 319 blocking effects of anti-IGF-I on CFU-E growth in the pres- the rejected grafts, and in the transplanted kidney in another ence of sera from ESRD patients with erythrocytosis and one who also had graft rejection. Erythrocytosis is a well anemia were not different (P ϭ 0.8972). The stimulatory known complication after renal transplantation, but it usually effects of sera from ESRD with erythrocytosis could partially occurred in patients with excellent allograft function (6,7). be nullified by the addition of anti-IGF-I in patients 1, 7, and Unlike the postrenal transplant erythrocytosis, these three pa- 8 (Table 4). The neutralizing effect of anti-IGF-I was nega- tients had erythrocytosis when they resumed dialysis therapy tively correlated with serum IGF-I level (r ϭϪ0.6667, P ϭ after redeveloping terminal renal failure following chronic 0.0710), whereas there was no such correlation in ESRD pa- rejection. For the four patients with non-ACDK, all had severe tients with anemia (r ϭϪ0.08571, P ϭ 0.8717). anemia at the initiation of hemodialysis, and they developed erythrocytosis during the course of maintenance dialysis. There Discussion were no known extrarenal causes of secondary erythrocytosis, All eight patients in the present series were male; four including cardiopulmonary disease and presence of abnormal patients had ACDK. The diagnosis of ACDK in our patients hemoglobin with high affinity for oxygen. Repeated determi- was established by renal ultrasonography and/or computed nations of arterial blood oxygen content precluded that hypoxia tomography, which showed multiple cysts in both native kid- played a role. Other investigators found that hepatitis or hep- neys or renal graft. Although ACDK is a common complica- atocellular carcinoma might be accompanied by an increase in tion in patients with ESRD on maintenance dialysis (23–25), hematocrit and EPO levels in patients with ESRD on dialysis erythrocytosis has rarely been described in patients with (27–31). Normal liver enzymes and ␣-fetoprotein levels as ACDK (10). For patients with ACDK, the cysts usually devel- well as normal liver sonograms excluded the possibility that oped in nontransplanted patients and rarely in the renal graft liver disease was the cause of erythrocytosis in our patients. (26). Three of our four patients with ACDK had prior renal Our patients with ACDK had elevated or normal EPO levels. transplants with their native kidneys retained. Multiple cysts These results were comparable to those of serum EPO concen- developed in the native kidneys in two patients after removal of trations in patients with secondary polycythemia of renal cystic

Table 4. Serum hormone levels and effects of sera from ESRD patients with erythrocytosis on normal bone marrow CFU-E growtha

Serum IGF-I % of EPO-Dependent (%) Anti-IGF-I Patient Serum EPO Level (U/L) ␮ CFU-E with 10% c Level ( g/L) Patient Serumb Neutralizing Test

ACDK 1 7.7 (7.1, 8.2)d 470.3 114 68 2 65.9 (44.9) 333.0 114 99 3 55.7 (21.5) 247.5 152 89 4 88.4 (41.9, 60.2) 249.8 156 112 Non-ACDK 5 12.5 (47.4, 82.1) 135.0e 145 104e 6 120.0 (298, 124) 202.5 126 96 7 9.1 (4.5, 4.3) 416.3 153 72 8 12.6 (8.5) 362.3 118 66 Patients (n ϭ 8) mean Ϯ SD 46.5 Ϯ 42.8 302.1 Ϯ 113.1 135 Ϯ 18% 88 Ϯ 17% range 7.7 to 120 135.0 to 470.3 114 to 156% 66 to 112% Control subjects normal subjects (n ϭ 40) (n ϭ 44) mean Ϯ SD 9.28 Ϯ 2.41 199.2 Ϯ 42.7 range 5.80 to 15.30 97.9 to 384.8 ESRD with anemia (n ϭ 105) (n ϭ 20) (n ϭ 11) (n ϭ 6) mean Ϯ SD 24.2 Ϯ 42.4 174.7 Ϯ 84.0 84 Ϯ 13% 90 Ϯ 16% range 4.1 to 240 73.1 to 337.5 26 to 106% 67.0 to 112.0%

a CFU-E, colony forming unit-erythroid; EPO, erythropoietin; IGF-I, insulin-like growth factor I; ACDK, acquired cystic disease of the kidney. b Expressed as a percentage of those with 10% normal serum. c Expressed as the percentage of CFU-E number with the addition of anti-IGF-I of that without anti-IGF-I in the presence of 10% uremic serum. d Numbers in parentheses indicate levels of subsequent measurements during follow-up. e Measured with an EPO level of 82.1 U/L. 320 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 315–322, 1999 disease or but without renal failure cance. In contrast, patients with ESRD and erythrocytosis had (32,33). Edmunds et al. found that the development of ACDK significantly higher serum IGF-I levels than those of healthy was associated with increased EPO levels in patients receiving subjects, or patients with ESRD and anemia. These findings regular hemodialysis, but no patient in their study had a he- suggest that IGF-I might play a role in the regulation of moglobin greater than 15 g/dl (34). Shalhoub et al. described erythrocytosis in our patients with ESRD and erythrocytosis, two patients with erythrocytosis and increased serum EPO especially in those who did not have elevated serum EPO levels accompanying secondary cyst formation in long-term values. Urena et al. also found that IGF-I stimulated erythro- hemodialysis (10). The stimulation of EPO production in poiesis in uremic patients with severe hyperparathyroidism ACDK may be attributed to local hypoxia of remnant kidney (51). They demonstrated that serum IGF-I concentrations were tissue caused by the pressure of cystic formation, which bears well correlated with hematocrit levels; however, unlike our a resemblance to the cysts in autosomal dominant polycystic patients, their patients had a low hematocrit level with a mean kidney disease (9,35). Increased EPO production under condi- of 28.1 Ϯ 1.7%. To prove that the elevation in IGF-I was tions of erythrocytosis in our patients also supports the possi- causally related to the erythrocytosis in our patients, we per- bility of disturbed oxygen sensor function in the diseased formed a neutralizing experiment by adding the anti-IGF-I to kidneys of chronic renal failure as suggested by Eckardt et al. the in vitro erythroid progenitor assay in the presence of (35). On the other hand, patients with ACDK and erythrocy- patients’ sera. Our results showed that anti-IGF-I could par- tosis were not necessarily associated with an increased serum tially abolish the erythropoietic action of serum on CFU-E EPO level as seen in patient 1. The results of previous studies growth. The correlation coefficient between serum IGF-I level on the correlation between hematocrit level and degree of and the blocking effect of anti-IGF-I in this small series was cystic disease or kidney size were conflicting (25,36). We and not so high, which was partly attributed to the addition of others found that the kidney sizes of ACDK might be reduced, exogenous EPO in the cultures and the presence of other normal, or enlarged (38). We also found that serum EPO levels possible inhibitors or stimulators in the uremic serum. did not correlate with the degree of renal cystic disease. The Erythropoietic inhibitors in sera from ESRD patients with serum EPO levels of non-ACDK patients were consistently anemia have been demonstrated by others (52–56). Our results normal in two, markedly increased in one, and fluctuated in also showed that sera from undialyzed patients with ESRD and another one on repeated measurements. The cause of increased anemia inhibited CFU-E formation in normal human marrow EPO secretion in patients 5 and 6 was not identified in the cultures. Some investigators found that uremic sera contained long-term follow-up. dialyzable inhibitors of in vitro erythropoiesis (52,54,56). The The present study confirmed that serum EPO levels vary effects of sera from uremic anemic patients on normal marrow widely in uremic anemic patients and EPO levels have no CFU-E growth were comparable before and after hemodialysis correlation with the hematocrit levels (39,40). We also found in the present study. In contrast, we found that sera from that there was no correlation between serum EPO levels and patients with ESRD and erythrocytosis produced a significant hematocrit values in ESRD patients with erythrocytosis. None stimulation of normal marrow CFU-E growth. These findings of our ESRD patients had EPO-independent BFU-E colonies showed that their sera did not have an inhibiting effect, but that are characteristic of polycythemia vera (41,42). In the pres- rather enhanced erythropoietic activity. Since the effect of sera ence of EPO in cultures, the numbers of BFU-E colonies in from ESRD with erythrocytosis was the net effect of stimula- patients with ESRD and erythrocytosis were significantly in- tion and inhibition, the increased erythropoietic effect in the creased compared with those of normal subjects. These findings sera of ESRD patients with an elevated EPO or IGF-I levels suggest that their erythroid progenitors exhibited an increased might have overcome the uremic inhibitors. sensitivity to EPO. However, ESRD patients with erythrocytosis In conclusion, we reported the clinical and laboratory results or anemia did not show a statistical difference in the BM BFU-E of eight patients with a rare combination of disease: ESRD and number, which made the increased sensitivity of BFU-E to EPO erythrocytosis. The present results suggest that IGF-I could be an important factor responsible for erythrocytosis unlikely. an important factor regulating erythropoiesis in patients with It has been demonstrated that IGF-I enhanced human ery- ESRD and erythrocytosis, especially in those who did not have throid progenitor growth in vitro and IGF-I could substitute for an increased EPO production. EPO as a stimulator of erythropoiesis (43–46). Brox et al. have shown that EPO-like factor from an anephric patient’s serum Acknowledgments with a hematocrit of 38 to 40% and a low-to-normal EPO level This work was supported by grants from the National Science was able to stimulate late erythropoiesis (47). This erythroid Council (Taiwan, Republic of China; NSC83-0412-B182-072) and cell-stimulating factor was later purified and characterized as CMRP434 from Chang Gung Memorial Hospital. We thank Dr. H. S. IGF-I by Congote et al. (48). Three of our patients did not have Wang and Mr. Y. C. Ou for radioimmunoassays. We also thank Ms. L. C. See for statistical analysis and Ms. Y. F. Wang for secretarial elevated EPO levels on serial assays; we thus measured the assistance. serum IGF-I levels in ESRD patients with erythrocytosis and anemia. Serum IGF-I levels in adult uremic anemic patients References determined by RIA have been found to be low (49,50). Our 1. Thorling EB: Paraneoplastic erythrocytosis and inappropriate results showed that serum IGF-I levels in ESRD and anemia erythropoietin production: A review. Scand J Haematol 9[Suppl were lower than normal but did not reach statistical signifi- 17]: 16–43, 1972 J Am Soc Nephrol 10: 315–322, 1999 IGF-I in Uremia and Erythrocytosis 321

2. Donati RM, Lange RD, Gallagher NI: Nephrogenic erythrocyto- 23. Grantham JJ, Levine E: Acquired cystic disease: Replacing one sis. Arch Intern Med 112: 960–965, 1963 kidney disease with another. Kidney Int 28: 99–105, 1985 3. Aderson ET, Walker BR: Polycystic kidney disease, polycythe- 24. Matson MA, Cohen EP: Acquired cystic kidney disease: Occur- mia and azotemia. JAMA 208: 2472–2473, 1969 rence, prevalence, and renal . Medicine 69: 217–226, 4. Martt JM, Sayman A, Neal MP: Polycythemia and hydronephro- 1990 sis. Ann Intern Med 54: 790–795, 1961 25. Ishikawa I: Acquired cystic disease: Mechanisms and manifes- 5. Luke RG, Kennedy AC, Barr SW, McDonald GA: Renal artery tations. Semin Nephrol 11: 671–684, 1991 stenosis hypertension and polycythemia. Br Med J 1: 164–166, 26. Chung WY, Nast CC, Ettenger RB, Danoviton GM, Ward HJ, 1965 Cohen AH: Acquired cystic disease in chronically rejected renal 6. Wickre CG, Norman DJ, Bennison A, Barry JM, Bennett WM: transplants. J Am Soc Nephrol 2: 1298–1301, 1992 Postrenal transplant erythrocytosis: A review of 53 patients. 27. Kolk-Vegter AJ, Bosch E, van Leeuwen AM: Influence of serum Kidney Int 23: 731–737, 1983 hepatitis on haemoglobin level in patients on regular haemodi- 7. Qunibi WY, Barri Y, Devol E, Al-Furayh O, Sheth K, Taher S: alysis. Lancet i: 526–528, 1971 Factors predictive of post-transplant erythrocytosis. Kidney Int 28. Simon P, Meyrier A, Tanquerel T, Ang KS: Improvement of 40: 1153–1159, 1991 anaemia in haemodialysed patients after viral or toxic hepatic 8. Fisher JW: Mechanism of the anemia of chronic renal failure. cytolysis. Br Med J 280: 892–894, 1980 25: 106–111, 1980 29. Brown S, Caro J, Erslev AJ, Murray TG: Spontaneous increase 9. Pavlovic-Kentera V, Clemons GK, Djukanovic L, Biljanovic- in erythropoietin and hematocrit value associated with transient Paunovic L: Erythropoietin and anemia in chronic renal failure. liver enzyme abnormalities in an anephric patient undergoing Exp Hematol 15: 785–789, 1987 hemodialysis. Am J Med 68: 280–284, 1980 10. Shalhoub RJ, Rajan U, Kim VV, Goldwasser E, Kark JA, An- 30. Gordon AS, Zanjani ED, Zalusky R: A possible mechanism for toniou LD: Erythrocytosis in patients on long-term hemodialysis. the erythropoiesis associated with hepatocellular carcinoma in Ann Intern Med 97: 686–690, 1982 man. Blood 35: 151–157, 1970 11. Muto S, Asano Y, Hosoda S, Shionoya S, Miura Y, Urabe A, 31. Kew MC, Fisher JW: Serum erythropoietin concentrations in Takaku F: Polycythemia of end-stage renal failure: No inhibition patients with hepatocellular carcinoma. 58: 2485–2488, 1986 of erythropoiesis by uremic serum and markedly increased serum 32. Koeffler HP, Goldwasser E: Erythropoietin radioimmunoassay in erythropoietin level. Nephron 46: 34–36, 1987 evaluating patients with polycythemia. Ann Intern Med 94: 44– 12. Chagnac A, Zevin D, Weinstein T, Gafter U, Korzets A, Levi J: 47, 1981 Erythrocytosis associated with renal artery thrombosis in a pa- 33. Murphy GP, Mirand EA, Johnston GS, Gibbons RP, Schirmer tient with polycystic kidney disease on hemodialysis. Acta HKA, Scott WW: Erythropoietin alterations in human genitouri- Haematol 84: 40–42, 1990 nary disease states: Correlation with experimental observations. 13. Bender FH, Piraino B: Polycythemia in diabetic patients on J Urol 99: 802–810, 1968 CAPD. Adv Perit Dial 7: 77–80, 1991 34. Edmunds ME, Devoy M, Tomson CRV, Krishna U, Clayworth 14. Spivak JL, Cooke CR: Polycythemia vera in an anephric man. A, Durrant STS, Feehally J, Walls J: Plasma erythropoietin levels Am J Med Sci 272: 339–344, 1976 and acquired cystic disease of the kidney in patients receiving 15. Jermanovich NB: Polycystic kidney disease and polycythemia regular haemodialysis treatment. Br J Haematol 78: 275–277, vera: Occurrence in a patient receiving hemodialysis. Arch Intern 1991 Med 143: 1822–1823, 1983 35. Eckardt K-U, Dru¨eke T, Leski M, Kurtz A: Unutilized reserves: 16. Bonet-Sol J, Flores-Lopez A, Caralps-Riera A: Polycythaemia The production capacity for erythropoietin appears to be con- vera: An exceptional cause of erythrocytosis in chronic renal served in chronic renal disease. Contrib Nephrol 88: 18–31, failure [Letter]. Nephron 59: 314, 1991 1991 17. Berlin NI: Diagnosis and classification of the . 36. Ishikawa I: Uremic acquired cystic disease of kidney. Urology Semin Hematol 12: 339–351, 1975 25: 101–108, 1985 18. Shih LY, Leu ML: Erythrocytosis in patients with renal failure 37. Deleted in proof. on hemodialysis: Study of underlying mechanism by in vitro 38. Fick GM, Gabow PA: Hereditary and acquired cystic disease of erythroid culture assay. Exp Hematol 21: 1239–1243, 1993 the kidney. Kidney Int 46: 951–964, 1994 19. Messner HA, Till JE, McCulloch EA: Interacting cell popula- 39. Radtke HW, Claussner A, Erbes PM, Scheuermann EH, Scho- tions affecting granulopoietic colony formation by normal and eppe W, Koch KM: Serum erythropoietin concentration in leukemic human marrow cells. Blood 42: 701–710, 1973 chronic renal failure: Relationship to degree of anemia and 20. Iscove NN, Sieber F, Winterhalter KH: Erythroid colony forma- excretory renal function. Blood 54: 877–884, 1979 tion in cultures of mouse and human bone marrow: Analysis of 40. De Klerk G, Wilmink JM, Rosengarten PCJ, Vet RJWM, the requirement for erythropoietin by gel and affinity Goudsmit R: Serum erythropoietin (ESF) titers in anemia of chromatography on agarose-concanavalin A. J Cell Physiol 83: chronic renal failure. J Lab Clin Med 100: 720–734, 1982 309–320, 1974 41. Prchal JF, Axelrad AA: Bone marrow response in polycythemia 21. Ogawa M, Grush OC, O’Dell RF, Hara H, MacEachern MD: vera [Letter]. N Engl J Med 290: 1382, 1974 Circulating erythropoietic precursors assessed in culture: Char- 42. Shih LY, Lee CT, Ou YC: Prediction of clinical course in acterization in normal men and patients with hemoglobinopa- patients with idiopathic erythrocytosis by endogenous erythroid thies. Blood 50: 1081–1092, 1977 colony assay but not by serum erythropoietin levels. Exp Hema- 22. Leu ML, Shih LY, Huang MJ: Serum ferritin and bone marrow tol 25: 288–292, 1997 iron in patients on maintenance hemodialysis. J Formosan Med 43. Akahane K, Tojo A, Urabe A, Takaku F: Pure erythropoietic Assoc 80: 804–814, 1981 colony and burst formations in serum-free culture and their 322 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 315–322, 1999

enhancement by insulin-like growth factor I. Exp Hematol 15: 50. Goldberg AC, Trivedi B, Delmez JA, Harter HR, Daughaday 797–802, 1987 WH: Uremia reduces serum insulin-like growth factor I, in- 44. Claustres M, Chatelain P, Sultan C: Insulin-like growth factor I creases insulin-like growth factor II, and modifies their serum stimulates human erythroid colony formation in vitro. J Clin binding. J Clin Endocrinol Metab 55: 1040–1045, 1982 Endocrinol Metab 65: 78–82, 1987 51. Urena P, Bonnardeaux A, Eckardt K-U, Kurtz A, Drueke TB: 45. Merchav S, Tatarsky I, Hochberg Z: Enhancement of erythro- Insulin-like growth factor I: A modulator of erythropoiesis in poiesis in vitro by human is mediated by insu- uraemic patients. Nephrol Dial Transplant 7: 40–44, 1992 lin-like growth factor I. Br J Haematol 70: 267–271, 1988 52. Ohno Y, Rege AB, Fisher JW, Barona J: Inhibitors of erythroid 46. Correa PN, Eskinazi D, Axelrad AA: Circulating erythroid pro- colony-forming cells (CFU-E and BFU-E) in sera of azotemic genitors in polycythemia vera are hypersensitive to insulin-like patients with anemia of renal disease. J Lab Clin Med 92: growth factor-I in vitro: Studies in an improved serum-free 916–923, 1978 medium. Blood 83: 99–112, 1994 53. Radtke HW, Rege AB, LaMarche MB, Bartos D, Bartos F, 47. Brox AG, Congote F, Fafard J, Fauser AA: Identification and Campbell RA, Fisher JW: Identification of spermine as an in- characterization of an 8-kD stimulating late erythropoi- hibitor of erythropoiesis in patients with chronic renal failure. esis. Exp Hematol 17: 769–773, 1989 J Clin Invest 67: 1623–1629, 1981 48. Congote LF, Brox A, Lin FK, Lu HS, Fauser AA: The N- 54. Summerfield GP, Bellingham AJ: The effects of therapeutic terminal sequence of the major erythropoietic factor of an aneph- dialysis and renal transplantation on uraemic serum inhibitors of ric patient is identical to insulin-like growth factor I. J Clin erythropoiesis in vitro. Br J Haematol 58: 295–304, 1984 Endocrinol Metab 72: 727–729, 1991 55. McGonigle RJS, Wallin JD, Shadduck RK, Fisher JW: Erythro- 49. Powell DR, Rosenfeld RG, Baker BK, Liu F, Hintz RL: Serum poietin deficiency and inhibition of erythropoiesis in renal insuf- levels in adults with chronic renal failure: The ficiency. Kidney Int 25: 437–444, 1984 importance of measuring insulin-like growth factor I (IGF-I) and 56. Delwiche F, Segal GM, Eschbach JW, Adamson JW: Hemato- IGF-II in acid-chromatographed uremic serum. J Clin Endocri- poietic inhibitors in chronic renal failure: Lack of in vitro spec- nol Metab 63: 1186–1192, 1986 ificity. Kidney Int 29: 641–648, 1986