Hypercalcemia and Overexpression of CYP27B1 in a Patient With Nephrogenic Systemic Fibrosis: Clinical Vignette and Literature Review

Authors

  • Vivian Y Pao,

    1. Endocrine-Metabolism Section, Department of Veterans Affairs Medical Center, University of California, San Francisco, California, USA
    Search for more papers by this author
  • Sandra Chang,

    1. Endocrine-Metabolism Section, Department of Veterans Affairs Medical Center, University of California, San Francisco, California, USA
    Search for more papers by this author
  • Dolores M Shoback,

    1. Endocrine-Metabolism Section, Department of Veterans Affairs Medical Center, University of California, San Francisco, California, USA
    Search for more papers by this author
  • Daniel D Bikle

    Corresponding author
    1. Endocrine-Metabolism Section, Department of Veterans Affairs Medical Center, University of California, San Francisco, California, USA
    • Address correspondence to: Daniel Bikle, MD, PhD, Endocrine Unit, 111N, VA Medical Center, 4150 Clement Street, San Francisco, CA 94121, USA
    Search for more papers by this author

  • The authors state that they have no conflicts of interest.

  • Published online on December 29, 2008

Abstract

Nephrogenic systemic fibrosis (NSF) is a disease of thickened, hard, hyperpigmented skin lesions with or without systemic fibrosis occurring in patients with renal insufficiency and associated with the administration of gadolinium-containing contrast. The pathogenesis of this disease is unclear, and there is no definitive treatment. We describe a 71-yr-old patient with stable chronic lymphocytic leukemia (CLL), end-stage renal disease (ESRD), and NSF who presented with hypercalcemia in 2006. Before onset of renal insufficiency in 2002, serum calcium, phosphorus, and PTH levels were normal. In 2004, the patient began hemodialysis, and he was diagnosed with NSF in 2005, shortly after undergoing an MRI with gadolinium contrast administration. Over the next 6 mo, albumin-corrected serum total calcium levels rose from 9.9 to 13.1 mg/dl (normal range, 8.5–10.5 mg/dl) with normal serum phosphorus levels. On admission in September 2006, 1,25-dihydroxyvitamin D [1,25(OH)2D] levels were elevated at 130.7 pg/ml (normal range, 25.1–66.1 pg/ml). Biopsy of an NSF lesion showed increased 25-hydroxyvitamin D3–1-α hydroxylase (CYP27B1) immunostaining compared with the biopsy from a normal control. This is the first reported association of NSF with hypercalcemia caused by elevated 1,25(OH)2D levels. This metabolic disturbance should be sought in future cases to determine a connection between NSF, 1,25(OH)2D metabolism, and CYP27B1 activation in the skin, which may shed light on the pathogenesis of this unusual local and systemic fibrosing disorder.

INTRODUCTION

Nephrogenic systemic fibrosis (NSF) is a disease seen in patients with renal disease that is characterized by thick, hardened, and hyperpigmented skin similar to scleroderma and/or systemic fibrosis.(1,2) A report linking NSF and gadolinium contrast led to an FDA advisory in 2006, warning against the use of gadolinium-based contrast agents in patients with kidney failure.(3)

We present a case of a patient with NSF, end-stage renal disease (ESRD), and hypercalcemia. Our patient had elevated 1,25-dihydroxyvitamin D [1,25(OH)2D] levels and increased 25-hydroxyvitamin D3–1-α hydroxylase (CYP27B1) expression in the NSF-affected skin lesions. Treatment with prednisone reduced his serum calcium and 1,25(OH)2D levels. To our knowledge, this is the first report of NSF accompanied by hypercalcemia.

CASE REPORT

A 71-yr-old man with ESRD, aortic stenosis, chronic lymphocytic leukemia (CLL), and NSF presented with delirium, weakness, vomiting, and constipation in September 2006. The medical history showed a diagnosis of CLL confirmed by lymph node biopsy in 1998. He completed multiple cycles of chemotherapy in 2004. On presentation in September 2006, an oncology consultant assessed the CLL as low grade and slowly progressive. Renal deterioration was noted in 2001 when serum creatinine rose from 1.1 to 5.0 mg/dl, and a kidney biopsy showed lymphomatous invasion of the renal parenchyma and tubular atrophy. In 2004, serum PTH levels rose to 378 pg/ml and stabilized between 207 and 255 pg/ml (normal range, 10–65 pg/ml) after the initiation of dialysis and calcitriol. Albumin-corrected serum total calcium levels remained in the normal range (Fig. 1). Serum phosphorus levels were between 3.4 and 5.9 mg/dl (normal range, 2.5–4.5 mg/dl). Notably, before the onset of renal insufficiency in 2001, corrected total serum calcium was 9.6 mg/dl (normal range, 8.5–10.5 mg/dl), with a normal serum PTH level.

Figure Figure 1.

Serum corrected total calcium, creatinine, and intact PTH levels from September 2001 to July 2006. Corrected calcium = 0.8 × (normal albumin – patient's albumin) + serum calcium. Normal albumin used was 4.0. Cr, creatinine; Ca, calcium; ESRD, end-stage renal disease; HD, hemodialysis; NSF, nephrogenic systemic fibrosis.

In September 2005, the patient underwent an MRI with gadolinium contrast for routine evaluation of CLL. Several weeks later, he noticed painful, thickened, and hyperpigmented plaques on all four extremities with “sausage” digits. Skin biopsy showed changes consistent with NSF. In November 2005, the vitamin A analog acitretin (25 mg orally daily) was started as treatment.

In February 2006, corrected total calcium was 12.1 mg/dl (Fig. 1). Intact PTH was 105.1 pg/ml. Serum PTH related-peptide (PTHrP) level was undetectable; serum phosphorus was 4.5 mg/dl. All vitamin D supplements were discontinued at this time.

In July 2006, corrected serum total calcium remained elevated at 12.4 mg/dl. Repeat PTH and PTHrP levels were unchanged, and serum phosphorus was 3.9 mg/dl. 25(OH)D level was 23.3 ng/ml (normal range, 10–60.0 ng/ml). Serum and urine electrophoreses did not show monoclonal spikes. Skeletal survey was negative for lytic lesions. A low calcium bath of 2 mg/dl (standard bath, 2.5 mg/dl) was initiated in hemodialysis, and pamidronate (20 mg, IV) was given in July 2006. One week later, the serum calcium level was unchanged.

In September 2006, weekly pamidronate (20 mg, IV) was initiated, and the patient was dialyzed with 1.5 mg/dl calcium baths three to four times weekly. Serum calcium improved only temporarily after dialysis. The patient was admitted on September 28, 2006 for daily hemodialysis with low calcium baths and further evaluation. On admission, his medications were acitretin, sertraline, trazodone, oxycodone, sevelamer, and quetiapine.

Physical examination showed an alert, oriented man with normal vital signs. Generalized lymphadenopathy was noted. Skin exam showed shiny, hyperpigmented, and woody induration involving bilateral wrists and hands, sparing the upper arms but involving the lower extremities from the toes to the distal thighs (Fig. 3A).

Admission laboratory studies included corrected serum total calcium of 13.0 mg/dl, phosphorus of 3.4 mg/dl, ionized calcium of 1.36 mM (normal range, 1.12–1.32 mM), and PTH of 99 pg/ml.

Daily hemodialysis with a low calcium (1.5 mg/dl) bath was initiated. The oncology consultant thought it unlikely that CLL was contributing to hypercalcemia in the absence of lytic lesions or evidence of acute worsening of his leukemia. Acitretin was discontinued. It was again verified that no vitamin D analogs were being administered to the patient at any time and specifically with dialysis. Despite these efforts, the patient continued to be hypercalcemic (Fig. 2). Repeat PTH level was 59.5 pg/ml, and free vitamin A level was 57 μg/dl (normal range, 38–106 μg/dl). The 1,25(OH)2D level was markedly elevated at 130 pg/ml (normal range, 25.1–66.1 pg/ml), which was particularly noteworthy for this ESRD patient who was on hemodialysis and not receiving any calcitriol. On October 7, 2006, pamidronate (40 mg, IV) was given without an effect on calcium levels. Four days later on October 11, 2006, prednisone (40 mg orally daily) was initiated to treat presumed 1,25(OH)2D-mediated hypercalcemia. On October 12, 2006, the 1,25(OH)2D level had fallen to 41.7 pg/ml. Ionized calcium levels also normalized. Two days after prednisone was initiated, dialysis therapy with low calcium baths (1.5 mg/dl) was reduced to four times weekly. The patient was discharged on October 16, 2006 taking prednisone (40 mg daily). Shortly after discharge while taking prednisone, the patient underwent biopsy of his NSF-affected skin.

Figure Figure 2.

Ionized calcium levels during hospitalization from September to October 2006. HD, hemodialysis; 1,25D, 1,25(OH)2D levels; 4x/wk, four times weekly.

A 3 × 1-cm sample of NSF-affected skin from the left lateral thigh was biopsied. A skin biopsy sample from a subject who did not have NSF was used for comparison. Paraffin-embedded tissues were cut in 5-μm sections. After deparaffinization and rehydration, the sections were quenched with hydrogen peroxide (3%) in PBS to block endogenous peroxidase activity. Tissues were incubated with BSA (4%) in Tris-buffered saline to block nonspecific binding. Sheep anti-murine 1-α hydroxylase antibody (Binding Site, Birmingham, UK) was applied to each section (1:150) and incubated overnight at 4°C. The specificity of this antibody has been confirmed by Western blot.(4) The primary antibody was detected with biotinylated rabbit anti-sheep anti-sera at 1:200 followed by ABC peroxidase reagent (both from Vector Laboratories, Burlingame, CA, USA). Enzyme activity was localized with diaminobenzidine substrate (Vector Laboratories, Burlingame, CA, USA). Omission of the primary antibody resulted in no signal, indicating the specificity of the immunodetection.

The skin biopsy (Fig. 3C) showed a uniform and marked increased expression of CYP27B1 throughout the epidermis extending into the basal layer in the NSF-affected skin lesion compared with the normal skin biopsy. A corresponding H&E stain (Fig. 3B) showed a paucity of lymphocytes in the skin. The few lymphocytes present were of normal morphology and clustered around blood vessels.

Figure Figure 3.

(A) Photograph of the patient's affected hand showing shiny, thickened, hyperpigmented, woody induration of skin and “sausage” digits. (B) Skin biopsy: H&E stain. The H&E stain of the patient's NSF-affected skin (from the same biopsy specimen as in C) shows a paucity of lymphocytes in the skin. The few lymphocytes present are of normal morphology and clustered around blood vessels. (C) Skin biopsies: CYP27B1 staining. Left panel is a biopsy of the patient's NSF-affected skin and shows uniform, increased staining of CYP27B1 throughout the entire epidermis extending down to the basal layer. Middle panel from a normal control patient indicates much less prominent staining for CYP27B1 in the epidermis. The right panel is a negative control that is stained without the primary antibody. The dark brown color in the basal layer is caused by pigmentation from melanocytes.

DISCUSSION

Hypercalcemia in a patient with ESRD such as this one suggests several etiologies including hypervitaminosis A caused by the vitamin A analog acitretin, malignancy (CLL), hypervitaminosis D, and coincidental primary hyperparathyroidism. Ingestion of vitamin D or its analogs was vigorously considered, but there was no record of such drug administration. Although intact PTH levels were never undetectable in this patient, the abrupt onset of hypercalcemia, the presence of multiple prior normal calcium values, the history of renal failure, and the response to glucocorticoids all made it unlikely that primary hyperparathyroidism was the cause of this patient's hypercalcemia. It was more likely that the mildly elevated PTH levels reflected the nonsuppressibility of chronic renal secondary hyperparathyroidism. There was no evidence for other etiologies of hypercalcemia (thyrotoxicosis, pheochromocytoma, granulomatous disease, or adrenal insufficiency). Several prior normal calcium levels also ruled out familial benign hypocalciuric hypercalcemia.

The possibility that CLL or another malignancy might be the etiology for this patient's hypercalcemia was considered. His CLL had had a chronic stable course and showed no evidence of progression. PTHrP levels were not elevated, and there was no evidence of bony lesions or Richter's transformation to lymphoma. Without these features, oncologic consultants thought it highly unlikely for CLL to cause hypercalcemia. Furthermore, the lack of a monoclonal spike on protein electrophoresis and lytic lesions on skeletal survey both argued against multiple myeloma. Vitamin A toxicity also was considered. Acitretin was discontinued without any effect on calcium levels, and the free vitamin A level was normal. Additionally, the markedly elevated 1,25(OH)2D level would not be explained by hypervitaminosis A.

Markedly elevated 1,25(OH)2D levels and increased staining for CYP27B1 in NSF-affected skin supports our hypothesis that increased CYP27B1 activity in his skin underlies the pathogenesis of hypercalcemia. Indeed, his ESRD along with lymphomatous infiltration of his kidneys strongly argues against a renal source for the elevated 1,25(OH)2D levels. We have no definitive proof, however, of the source(s) for the 1,25(OH)2D levels in the blood. Successful treatment with prednisone, an agent known to suppress ectopic 1,25(OH)2D production(5) and block 1,25(OH)2D action in the gut,(6) resulted in improved calcium and 1,25(OH)2D levels in this patient, highlighting the critical role of 1,25(OH)2D. In addition, when glucocorticoids were tapered during subsequent outpatient visits, the patient's serum calcium levels rose to >11 mg/dl. When prednisone was increased to 20 mg orally daily, his serum calcium levels normalized. The patient was maintained on prednisone until his death in September 2007, from complications of bowel infarction and ESRD.

The pathogenesis of NSF remains uncertain. NSF is a fibrosing disorder in patients with renal insufficiency, both acute and chronic, and not always associated with hemodialysis.(1,7) Systemic fibrosis involving the heart, lung, esophagus, diaphragm, and other organs have been seen at autopsy.(2,8)

It is hypothesized that gadolinium-based contrast agents induce or contribute to NSF in patients with renal insufficiency through delayed clearance of contrast media caused by impaired renal function.(9) This prolonged exposure to gadolinium ions allows them to dissociate, diffuse into tissues, and cause damage.(10) In patients with normal renal function, the half-life of gadolinium is ∼2 h. In patients with ESRD, this extends to >5 h.(3,11)

No effective treatment for NSF exists, aside from recovery of renal function in acute renal insufficiency or renal transplantation in ESRD.(12) NSF is reported to increase mortality of patients on dialysis.(13)

This patient showed increased expression of CYP27B1 in NSF-affected skin compared with the skin of a normal control. Extrarenal expression of CYP27B1 is described in skin, lymph nodes (granulomata), colonic epithelial cells, pancreatic islet cells, adrenal medulla, brain, and placenta.(14,15) Enzyme activity at these sites is thought to provide 1,25(OH)2D for tissue-specific autocrine or paracrine functions.(14) In normal skin, CYP27B1 is expressed in keratinocytes and hair follicles.(15) TNF-α and IFN-γ have been shown to stimulate 1,25(OH)2D production in human keratinocytes.(16,17) Increased CYP27B1 expression in the skin is reported in psoriasis and sarcoidosis.(14) In psoriatic skin, there is widespread expression of CYP27B1 throughout the entire stratum spinosum in a dysregulated fashion.(14) However, psoriasis does not cause hypercalcemia. Interestingly, one patient with granulomatous “slack skin” secondary to a T-cell lymphoproliferative disorder showed hypercalcemia and elevated serum 1,25(OH)2D levels.(18) Skin biopsy showed overexpression of CYP27B1 in dermal granulomata and basal epidermal cells in affected skin, whereas uninvolved skin showed normal expression of CYP27B1 in the basal epidermis.(18) The hypercalcemia was thought to be caused by increased CYP27B1 activity in the immune cell population of the dermis. Macrophages lack significant 24-hydroxylase (CYP24) activity, the enzyme that converts 1,25(OH)2D into less active metabolites. In contrast, epithelial cells are capable of upregulating CYP24 by 1,25(OH)2D3 and therefore minimize the amount of 1,25(OH)2D reaching the circulation.(18)

Lesions in NSF contain increased collagen and dermal mucin, prominent collagen bundles, and greater numbers of dermal spindle-shaped cells.(19) NSF is thought to be mediated by circulating fibrocytes (CFs), cells that produce the histologic changes in NSF.(20) These cells are involved in normal wound healing and scar formation and are capable of producing the pathologic fibrosis seen in NSF.(10) It has been suggested that CFs may target or be stimulated by gadolinium deposition in the skin.(21) It may be that CFs also lack the ability to upregulate CYP24 activity to minimize 1,25(OH)2D levels, but this has not been addressed.

We propose that hypercalcemia is a novel manifestation of NSF and suggest that increased CYP27B1 activity in the NSF-affected skin lesions is the basis for this patient's hypercalcemia. Although CLL is difficult to exclude completely, it was stable during the hypercalcemic episode and was unlikely to be the cause of hypercalcemia. Prednisone treatment to antagonize production and action of 1,25(OH)2D resulted in a fall in both ionized calcium and 1,25(OH)2D levels. We speculate that the fall in 1,25(OH)2D levels was caused by suppression of cytokines that drive the activity but not the expression of the enzyme,(16,17) because overexpression of CYP27B1 persisted in the skin after prednisone therapy, when his circulating 1,25(OH)2D levels had fallen. Because NSF increasingly is being recognized, we recommend careful monitoring of patients with NSF for metabolic complications, especially hypercalcemia.

Acknowledgements

We thank Drs. Martin Hewison, Maria Serrano, and Arnaud Teichert for assistance.

Ancillary