Heterogeneity of Cultured Melanocyte Elongation and Proliferation Factor in Bilateral Diffuse Uveal Melanocytic Proliferation

Abstract

A patient with bilateral diffuse uveal melanocytic proliferation (BDUMP) associated with endometrial cancer was treated with plasmapheresis, but failed therapy with progressive serous retinal detachment. We collected plasma before and after plasmapheresis therapy. Our goal was to determine if the cultured melanocyte elongation and proliferation (CMEP) factor and hepatocyte growth factor (HGF) was present in the IgG enriched fraction and understand why our patient failed plasmapheresis therapy. Melanocytes were cultured for 3-5 days in the presence of control medium, unfractionated pre-plasmapheresis BDUMP medium, IgG enriched or IgG depleted BDUMP medium, or unfractionated post-plasmapheresis BDUMP medium. Subretinal fluid was collected from patients with BDUMP and control retinal detachments and analyzed by electropheresis with immunoblotting. Medium with unfractionated BDUMP plasma stimulated melanocyte growth 1.4-1.5 fold compared to control medium on days 3-5 (p < 0.001 for all). Both IgG enriched and IgG depleted BDUMP medium mildly increased melanocyte growth 1.3 fold (p < 0.05 for enriched, p < 0.01 for depleted) compared to control. In comparison, unfractionated BDUMP medium caused a 1.7-fold increase in melanocyte growth, which was significantly more than the enriched (p < 0.01) and depleted (p < 0.05) fractions. Pre-plasmapheresis and post-plasmapheresis unfractionated BDUMP medium equally stimulated melanocyte growth 1.7-fold (p < 0.05) compared to control. HGF was present in IgG depleted, pre-plasmapheresis, and post-plasmapheresis samples, but absent in the IgG enriched fraction. There was no enrichment of IgG in the subretinal fluid from eyes with BDUMP. In conclusion, CMEP factor is not concentrated in the IgG enriched plasma fraction in our patient who failed plasmapheresis therapy. HGF levels have no correlation with melanocyte growth. Because plasmapheresis preferentially removes immunoglobulins from the plasma, our patient responded poorly to plasmapheresis treatment with worsening retinal detachment.

(3). Introduction

Bilateral diffuse uveal melanocytic proliferation (BDUMP) is a rare, paraneoplastic disease. Due to increased disease awareness and longer survival of patients with cancer, the incidence of BDUMP has increased from 1.15 cases per year from 1980 – 2000 to 4.4 cases per year during the last 5 years (Klemp et al., 2017). The cardinal features of BDUMP reported by Gass include: (1) multiple round or oval red patches in the posterior fundus at the level of the RPE, (2) early hyperfluorescence on fluorescein angiography correlating with these patches, (3) diffuse uveal tract thickening with multiple slightly elevated pigmented or non-pigmented uveal tumors, (4) serous retinal detachment (SRD), and (5) worsening cataracts (Gass et al., 1990).

Plasmapheresis has recently emerged as a potential treatment for BDUMP. Miles et al hypothesized that an unknown circulating factor, termed cultured melanocyte elongation and proliferation factor (CMEP), stimulates melanocyte growth. This CMEP factor was also hypothesized to be either made and secreted by the cancer cells or produced by the immune system in response to the cancer (Miles et al., 2012). Miles et al demonstrated that serum from two patients with ovarian carcinoma and BDUMP increased melanocyte growth, and that CMEP factor was present in the IgG enriched and not in the IgG depleted fraction (Miles et al., 2012). Additionally, they showed that CMEP factor did not exist in the serum from other patients with paraneoplastic syndromes, and this effect was specific to BDUMP (Miles et al., 2012). However, no further characterization of CMEP has been performed. Based upon these data, plasmapheresis has been shown to improve vision and SRF (subretinal fluid) in 5 patients with BDUMP (Jansen et al., 2015; Mets et al., 2011; Pulido et al., 2013; Schelvergem et al., 2015).

Here, we report a case of BDUMP associated with clear cell endometrial carcinoma. Unlike prior case reports, the SRD progressed despite plasmapheresis therapy. We collected plasma from the 1^st and final plasmapheresis sessions to determine if the BDUMP plasma stimulated melanocyte growth, if CMEP was in the IgG enriched fraction in our patient, and if that factor was still present after the final plasmapheresis session.

(4). Materials and Methods

(4.1). Plasmapheresis

This work adhered to the tenets of the Declaration of Helsinki, is HIPAA compliant, informed consent was obtained from the subjects after explanation of the nature and possible consequences of the study, and was approved by the institutional review board at the Cleveland Clinic. Plasmapheresis was performed at the Cleveland Clinic. Over a period of 9 days, 6 sessions of plasmapheresis were performed using 500 mL of 5% human serum albumin as replacement serum. Plasmapheresis samples were collected from the 1^st session, the final session, and frozen at −80°C.

(4.2). Cell Culture Experiments

Primary human melanocyte cells (HEMnLP, Cat #C0025C) were purchased from Thermo Fisher Scientific (Waltham, MA). Experiments were performed on Passage 3-9. Cells were cultured in 1% penicillin-streptomycin (Pen/Strep, Cat #725-100, Lerner Research Institute Media Core, Cleveland, OH), 1% Human Melanocyte Growth Serum (HMGS, Cat #50025, Thermo Fisher Scientific), and 254 Medium (Cat #M254500, Thermo Fisher Scientific). See Table 1 for details about each growth medium. Cells were split into a 12 well plate and seeded overnight. Cells were cultured for 3-5 days with daily media changes. For cell counting, cells were trypsinized for 5 minutes and transferred to 1.7 mL tubes. Cells were centrifuged at 500 × g for 5 minutes and resuspended in 300 pL of growth medium. Cells were counted by hemacytometer.

Table 1.

Description of human melanocyte growth media. Parentheses indicate final concentration of added plasma fractions.

	Growth	Control	BDUMP	IgG Depleted	IgG Enriched
254 Medium	98%	98%	98%	98%	98%
Pen/Strep	1%	1%	1%	1%	1%
HMGS	1%	0.5%	0.5%	0.5%	0.5%
BDUMP			0.5% (0.43 mg/ml)
IgG Enriched					0.5% (0.15 mg/ml)
IgG Depleted				0.5% (0.15 mg/ml)
PBS		0.5%

(4.3). IgG Enrichment

Plasma from the 1^st plasmapheresis session was fractionated into IgG depleted and IgG enriched samples using column affinity purification. Plasma was diluted 1:1 in Binding Buffer (PBS pH 8.0 without MgCl₂ and CaCl₂). Two NAb Protein G Spin Columns (Cat #89961, Thermo Fisher Scientific) were run in parallel. Columns were equilibrated with Binding Buffer and loaded with 10 mL of diluted plasma. Columns were incubated at room temperature with end-over-end mixing for 10 minutes to allow protein binding. The unbound fraction was removed by centrifugation at 1000 × g for 1 minute. The unbound fraction was saved as the flow through or IgG depleted fraction. The column was washed 3 times with Binding Buffer. Before elution, Neutralization Buffer (1M Tris-HCl pH 8.0) was added to collection tubes so that it was 10% of total elution volume. The column was eluted 4 times with Elution Buffer (0.1M glycine, pH 2.5). The initial 2 IgG enriched elutions from both columns were combined for maximum yield. The combined elution (IgG enriched) and unbound fraction (IgG depleted) were concentrated using Amicon Ultra Centrifugal Filtration Units (Cat #UFC900308, Millipore, Burlington, MA) and centrifugation at 5000 × g at 4°C for 1 hour. The IgG enriched sample was dialyzed using the Slide-A-Lyzer Cassette (Cat #66380, Thermo Fisher Scientific) in Dialysis Buffer (PBS with 0.9 mM CaCl₂ and 0.9mM MgCl₂, pH 7.4) for 12 hours at 4°C to remove elution salts.

(4.4). Subretinal Fluid Collection

From the BDUMP patient and the uveitic SRD, intraocular drainage of the subretinal fluid was performed after core vitrectomy and immediately after making a drainage retinotomy. The soft tip cannula was positioned at the retinotomy and manual aspiration was performed with the infusion cannula clamped for removal of concentrated subretinal fluid. For the rhegmatogenous retinal detachment, subretinal fluid drainage was performed similarly, but instead of a drainage retinotomy, the fluid was drained from the causative retinal break. For the multiple myeloma-associated SRD, external needle drainage was performed. Briefly, after core vitrectomy was completed, perfluorocarbon liquid was instilled to flatten the retina while a 27 gauge needle was placed through the sclera and choroid into the subretinal space under direct visualization. Subretinal fluid was manually aspirated while perfluorocarbon was instilled. All subretinal fluid samples were saved at −80°C.

(4.5). Electropheresis and Immunoblotting

Total protein was determined by BCA protein assay (Cat #23227, Thermo Fisher Scientific). Using equal total protein, samples were separated on 12% SDS-PAGE gels or non-reducing gels for hepatocyte growth factor (HGF) immunoblotting. Brilliant blue staining (Cat #24592, Thermo Fisher Scientific) was performed overnight at room temperature, and gels were then washed in ddH₂0 to remove background. Gelswere imaged using a GS-800 Calibrated Densitometer (Cat #170-7980, Bio-Rad, Hercules, CA) and the MagicScan V6.0 software (Cat #UMXCMW600D, Umax Technologies, Dallas, TX).

For immunoblotting, samples were transferred overnight at 4°C to polyvinylidene difluoride membranes. For IgG probing, membranes were washed with TBS-T buffer (20 mM Tris, 500 mM NaCl, pH 7.5, .01% Tween-20), blocked with 5% BSA in TBS-T for 1 hour, and stained with AffiniPure Goat Anti-Human IgG, Fcγ Fragment Specific antibody conjugated to horseradish peroxidase in 1% BSA in TBS-T for 30 minutes at room temperature (1:50,000, Cat #109-035-098, Jackson ImmunoResearch Laboratories, West Grove, PA). Blots were developed using Western Lightning Enhanced Chemiluminescence Substrate Pro (Cat #NEL121001EA, PerkinElmer, Waltham, MA).

Prior to HGF probing, membranes were stained according to the REVERT total protein stain protocol (#926-11016, Li-Cor Biosciences, Lincoln, NE). Membranes were then washed with ddH₂O and blocked with Odyssey Blocking Buffer (TBS) (#927-50000, Li-Cor Biosciences, Lincoln, NE). Membranes were stained with mouse monoclonal anti-human HGF antibody (1 ug/mL, #MAB294, R&D Systems, Minneapolis, MN) in a 1:1 mixture of TBS-T buffer and Odyssey Blocking Buffer (TBS) overnight at 4°C, followed by a 1 hour incubation at room temperature of IRDye 800CW goat anti-mouse IgG (H+L) (1:15,000, #925-32210, Li-Cor Biosciences, Lincoln, NE) in the same 1: 1 buffer mixture as the primary. Membrane were imaged using an Odyssey CLx Imaging System (#9140, Li-Cor Biosciences, Lincoln, NE) and the Image Studio V5.2 software (#9140-500, Li-Cor Biosciences, Lincoln NE).

(4.6). Statistical Analysis

Samples from each melanocyte growth experiment were run in triplicate on a given day and averaged. The experiments were replicated 3-5 times. Analysis was done in a paired fashion to mitigate variability in well seeding. Repeated measures ANOVA with Tukey’s multiple comparison post test was performed.

(5). Results

A 65 year old female presented to the Cole Eye Institute with blurred vision in both eyes of 2 months duration. Past medical history included Stage IIIC2, Grade 3 clear cell endometrial cancer that was diagnosed 2.5 months prior to presentation. The patient had surgical resection, but was currently not undergoing adjuvant therapy. Past ocular history included rapid cataract development, necessiting cataract surgery 2 months prior to presentation. One week prior to presentation, the patient was found to have retinal striae on fundus photography, several pigmentary lesions in the inferior temporal macula in the left eye (Fig 1B), and retinal pigment epithelial (RPE) thickening and subretinal (SRF) fluid on optical coherence tomography (OCT) in both eyes (Fig 1C–D). The patient was initially diagnosed with Harada’s disease and started on 80 mg of oral prednisone (>1 mg/kg) for 1 week duration.

Figure 1.

Imaging upon initial presentation. Fundus photography of the right (A) and left (B) eyes demonstrated retinal striae and inferior temporal pigmentary lesions (yellow box) in the left eye. OCT showed RPE thickening (blue arrow) and SRF (green pound sign).

Upon presentation to Cole Eye Institute 1 week later, the patient demonstrated hand motion vision, no anterior segment inflammation, pseudophakia, trace anterior vitreous cells, and bilateral inferior SRD (Fig 2A & 3A). OCT showed large bilateral bullous sub-macular fluid (Fig 2D & 3D), demonstrating clear progression on 1 week of prednisone therapy. Ultra-widefield fluorescein angiography (UWF-FA) displayed early hyperfluorescence, disc leakage, and late leakage of the early hyperfluorescent areas throughout the mid-periphery in both eyes (Fig 2B–C & 3B–C). Ultra-widefield indocyanine green angiography (UWF-ICGA) demonstrated hypocyanescent lesions in both eyes (Fig 2E–F & 3E). The hypocyanescent lesions were barely detectable in the right eye as a pigmentary change (Fig 2A, white arrows), not detectable in the left eye (Fig 3A), and demonstrated no abnormalities on the UWF-FA. B-scan ultrasonography identified diffuse choroidal thickening with no mass lesions in either eye.

Figure 2.

Multimodal imaging of the right eye. Ultra-widefield fundus photography (A) identified inferior SRD (blue arrows) and subtle superior temporal melanocytic proliferation (white arrows). UW-FA (B-C) demonstrated many expanding areas of hyperfluorescence and no findings in the region of the melanocytic proliferation (red asterisk). Vertical OCT raster scan through the macula (D) displayed bullous subretinal fluid (green pound sign). UW-ICGA (E-F) showed hypocyanescence corresponding to a large (red asterisk) and small (red arrow) choroidal melanocytic proliferation.

Figure 3.

Multimodal imaging of the left eye. Ultra-widefield fundus photography (A) identified inferior SRD (blue arrows) and subtle superior nasal pigmentary changes (red asterisk). UW-FA (B-C) demonstrated multiple areas of expanding hyperfluorescence and no findings in the region of the melanocytic proliferation (red asterisk). Vertical OCT raster scan through the macula (D) displayed bullous subretinal fluid (green pound sign). Late phase UW-ICGA (E) showed hypocyanescence corresponding to a large area of melanocytic proliferation (red asterisk). Ultra-widefield fundus photography after 6 plasmapheresis sessions (F) illustrated significant worsening of the SRD.

Based upon the pigmentary lesions upon initial presentation (Fig 1B), RPE thickening on OCT (Fig 1C–D), early hyperfluorescence on UWF-FA (Fig 2B & 3B), hypocyanescent lesions on UWF-ICGA (Fig 2F & 3E), diffuse uveal thickening on B-scan, recent endometrial cancer diagnosis, rapid cataract development, and progression despite high dose oral steroids for 1 week, the patient was diagnosed with BDUMP and prednisone thereapy was discontinued without taper. Recently, plasmapheresis has been reported as a successful therapy for BDUMP (Jansen et al., 2015; Mets et al., 2011; Pulido et al., 2013; Schelvergem et al., 2015), based upon the hypothesis that plasmapheresis removes the CMEP factor (Miles et al., 2012). Since poor vision was the patient’s primary rationale for refusing further cancer therapy, the patient elected to undergo RD repair in the right eye with silicone oil tamponade prior to plasmapheresis. After successful retinal re-attachment, the patient underwent 6 plasmapheresis treatments over a 9 day period. Despite plasmapheresis, the SRD significantly worsened in the left eye (Fig 2F), requiring surgical drainage of the subretinal fluid.

In order to understand why plasmapheresis failed to improve the SRD, we saved plasmapheresis samples from the first and final plasmapheresis treatments to test the ability of BDUMP plasma to stimulate primary human melanocyte growth. We found that BDUMP medium (0.5% BDUMP / 0.5% HMGS) significantly increased the number of melanocytes by 1.4 to 1.5-fold compared to control (0.5% HMGS) medium at Day 3, Day 4, and Day 5 (Fig 4A–C, p < 0.001 for all). Growth medium alone (1% HMGS) stimulated modest increases in melanocyte cell numbers by 1.2-fold at Day 3 (Fig 4A, p < 0.05) and 1.3 fold at Day 5 (Fig 4C, p < 0.01) compared to control (0.5% HMGS) medium.

Figure 4.

BDUMP plasma increased melanocyte growth. Melanocytes were cultured for 3 days (A), 4 days (B), and 5 days (C) in the presence of 1% HMGS growth medium, 0.5% HMGS control medium, and 0.5% BDUMP medium with 0.5% HMGS. Data were normalized to control medium for each day. BDUMP medium increased melanocyte growth at Day 3, Day 4, and Day 5 (N = 5). * p < 0.05, ** p < 0.01, *** p < 0.001.

Since it was previously reported that the CMEP factor is present in the IgG enriched plasma fraction(Miles et al., 2012), we similarly fractionated the BDUMP plasma into IgG depleted and IgG enriched fractions. After the BDUMP plasma passed through an IgG-binding protein G column, electrophoresis analysis demonstrated that the two largest components of the elution were IgG heavy chain at 50 kDa and IgG light chain at 25 kDa (Fig 5A). Additionally, the flow through displayed a relative depletion of the IgG heavy and light chain bands, compared to the unfractionated BDUMP plasma (Fig 5A). To confirm that the elution was enriched with IgG protein, an immunoblot was performed against the IgG heavy chain. Both the unfractionated BDUMP plasma and the elution contained significant amounts of IgG heavy chain, while IgG heavy chain was barely detectable in the flow through (Fig 5B). These data confirm that the elution was IgG enriched and the flow through was IgG depleted.

Figure 5.

CMEP factor was not enriched in the IgG fraction. Electrophoresis (A) of unfractionated plasma, flow through (FT), and elution showed increased abundance of the IgG heavy chain (IgG HC) 50 kDa band and the IgG light chain (IgG LC) 25 kDa band in the elution and decreased abundance of these bands in the flow through. Immunoblot analysis for IgG heavy chain (B) demonstrated large amounts of IgG heavy chain in the elution and unfractionated plasma with barely detectable IgG heavy chain in the flow through. (C) Melanocytes were cultured for 4 days in the presence of control 0.5% HMGS medium, positive control unfractionated BDUMP plasma medium, IgG depleted [IgG (−)] medium, and IgG enriched [IgG (+)] medium. Unfractionated BDUMP medium stimulated the most melanocyte growth. IgG depleted and enriched medium mildly increased melanocyte growth (N=5). * p < 0.05, ** p < 0.01, *** p < 0.001 vs control 0.5%. # p < 0.05, ## p < 0.01 vs unfractionated BDUMP.

To determine whether the CMEP factor was present in the IgG enriched or depleted fraction, IgG enriched and IgG depleted media were made with equal amounts of total protein, which contained less total protein than unfractionated BDUMP plasma (Table 1). We next tested the ability of IgG enriched and depleted media to stimulate melanocyte growth after 4 days in culture. IgG depleted and enriched media increased melanocyte growth 1.4-fold (Fig 5C, p < 0.01) and 1.3-fold (Fig 5C, p < 0.05) compared to control 0.5% HMGS media. Unfractioned BDUMP media stimulated melanocyte growth 1.7-fold compared to control (Fig 5C, p < 0.001), which was signifantly more than IgG depleted (p < 0.05) and IgG enriched (p < 0.01) media. These data suggest that the CMEP factor is not specifically enriched in the IgG fraction, but that both fractions contained growth enhancing factors.

Since our patient failed plasmapheresis therapy, we hypothesized that BDUMP plasma obtained after the final plasmapheresis session would still increase melanocyte growth. Electropheresis analysis of the pre- and post-plasmapheresis BDUMP plasma demonstrated that the IgG heavy and light chains were mildly reduced after 6 plasmapheresis sessions (Fig 6A). As expected, both pre- and post-plasmapheresis BDUMP plasma stimulated melanocyte growth by 1.7-fold compared to control (p<0.05 for both, Fig 6B).

Figure 6.

BDUMP plasma after plasmapheresis stimulates melanocyte growth. Electrophersis (A) of pre- and post-plasmapheresis BDUMP plasma showed less IgG heavy and light chain after 6 plasmapheresis sessions. (B) Melanocytes were cultured for 4 days in the presence of control 0.5% HMGS medium, pre-plasmapheresis BDUMP medium, and post-plasmapheresis BDUMP medium. Both pre- and post-plasmaphersis BDUMP medium stimulated equal melanocyte growth (N=3, * p < 0.05 vs control).

Recently, BDUMP was associated with high HGF levels and anti-retinal autoantibodies that cross-reacted with HGF(Niffenegger et al., 2018). This report suggests that HGF and anti-retinal autoantibodies to HGF may drive choroidal nevi growth and RPE damage in BDUMP. To investigate this hypothesis, we performed immunoblot analysis on our samples for HGF. We found that HGF was present in the IgG depleted, pre-plasmapheresis, and post-plasmapheresis samples (Fig 7B). Additionally, HGF was nearly absent in the IgG enriched sample (Fig 7B). Since all of these samples caused melanocyte growth, there was no correlation between HGF level and melanocyte growth.

Figure 7.

(A) Non-reducing gel showing IgG enrichment. (B) Immunoblot for HGF shows HGF is present in IgG depleted, pre-plasmapheresis, and post-plasmpheresis samples, and nearly absent in the IgG enriched fraction.

To help understand the mechanism of SRD in BDUMP, we collected subretinal fluid from each eye of our patient. For comparison, we saved SRF from additional patients with rhegmatogenous retinal detachment (RRD), SRD from multiple myeloma, and SRD associated with idiopathic uveitis. We performed both protein electropheresis and immunoblot analysis for IgG heavy chain with equal total protein loaded from all samples. We found that subretinal fluid from all patients had qualitatively equal levels of albumin (Fig 8A). Eyes from BDUMP and SRD from idiopathic uveitis had slightly increased levels of IgG heavy and light chain compared to multiple myeloma SRD and RRD (Fig 8A–B). We identified no significant enrichment of IgG levels in subretinal fluid from BDUMP eyes compared to controls.

Figure 8.

BDUMP subretinal fluid demonstrated no IgG enrichment. Electropheresis (A) of subretinal fluid from BDUMP, two SRDs, and one rhegmatogenous RD showed similar amounts of albumin, IgG heavy chain (IgG HC) and IgG light chain (IgG LC). Immunoblot (B) analysis of subretinal fluid for IgG HC displayed similar amounts of IgG HC in each sample.

(6). Discussion

In this report, we identified a patient with BDUMP secondary to clear cell endometrial carcinoma who presented with rapid cataract progression, pigmentary lesions (Fig 1B), RPE thickening (Fig 1C–D), early hyperfluorescence on UWF-FA (Fig 2B & 3B), diffuse uveal thickening, and bilateral SRDs (Fig 2–3), meeting 5 of 5 critera for BDUMP as originally described by Gass (Gass et al., 1990). Alternative diagnoses include uveitic SRD, or neoplastic choroidal infiltration from the endometrial cancer, causing SRD. Our patient did not respond to high dose oral steroid (>1 mg/kg) for 1 week of therapy and no rapid enlargement of the choroidal thickening or uveal tumors was observed during her care. These observations make uveitic and neoplastic SRD less likely, and favor a diagnosis of BDUMP.

Despite plasmapheresis treatment, the SRD of the left eye progressed, necessitating surgical SRD repair (Fig 2F). There are 10 published reports of plasmapheresis treatment for BDUMP. Six reports demonstrated improvement in vision or SRF with plasmapheresis treatment (Jaben et al., 2011; Mets et al., 2011; Niffenegger et al., 2018; Pulido et al., 2013; Schelvergem et al., 2015); four of 6 patients were diagnosed with lung cancer (1 ovarian) and 4 of 6 were undergoing concomitant chemotherapy. Among the remaining 5 cases, 3 reports showed stability (Alasil et al., 2017; Jaben et al., 2011) and 2 reports displayed worsening of SRD despite plasmapheresis treatment (Alrashidi et al., 2014; Navajas et al., 2011). Among these 5 poorly responding patients, 2 patients were diagnosed with ovarian cancer (both stable), 2 with uterine cancer (1 stable, 1 progressed), and 1 with no identifiable primary malignancy (progressed). Additionally, 3 of 5 patients were not undergoing chemotherapy. Therefore, uterine carcinoma, and lack of chemotherapy may be risk factors for a poor clinical response to plasmapheresis.

Despite the worsening SRD, the collected plasma samples did contain the CMEP factor, which stimulated melanocyte growth (Fig 4). Therefore, plasmapheresis removed the CMEP factor from the plasma. One possible explanation for poor clinical response in this patient is rebound of CMEP factor in the plasma. Rebound is determined by molecular weight, distribution between the intravascular and extravascular compartments, and the production of the factor (Williams and Balogun, 2014). Large molecular weight, intravascular location, and slow production all favor more efficient plasmapheresis removal. For example, complement components and coagulation factors rebound to preplasmapheresis levels after 72 hours due to small molecular weight and fast production, while IgG rebounds 2 weeks after treatment (Williams and Balogun, 2014). The CMEP factor was originally hypothesized by Miles et al to either be a molecule released by the cancer cells or produced by the immune system (Miles et al., 2012). Therefore, we hypothesize that the CMEP factor in our case of endometrial cancer is more likely to be a growth factor produced by the tumor, which demonstrates fast rebound with plasmapheresis. A tumor-secreted growth factor would be constantly produced (if not undergoing chemotherapy like our case), exist in the extravascular space, and be smaller in size than IgG, explaining the fast rebound of CMEP, and inefficient removal by plasmapheresis.

Confirming our hypothesis, IgG enrichment did not increase melanocyte growth (Fig 5C). The IgG enriched fraction contained far more IgG heavy and light chains than unfractionated BDUMP plasma (Fig 5A–B). If CMEP were an IgG, then the IgG enriched fraction should stimulate equal or more melanocyte growth than the unfractionated BDUMP plasma. Instead modest growth was found for both the IgG depleted and enriched fraction (Fig 5C). This likely occurred because the enrichment process is not a purification of IgG alone and other proteins were still present in the IgG enriched fraction (Fig 5A). Because the enriched and depleted media contained approximately 1/3 of the total protein compared to unfractionated BDUMP plasma (Table 1), reduced melanocyte growth was observed as CMEP was diluted in each fraction.

To further confirm our hypothesis that CMEP was a tumor-secreted factor, we next tested BDUMP plasma from the final plasmapheresis session. We found that postplasmapheresis plasma had modestly reduced IgG heavy and light chains, yet stimulated equal melanocyte growth compared to pre-plasmapheresis plasma (Fig 6). Therefore, after 6 plasmapheresis sessions over 9 days, CMEP was present at equal abundance, supporting our hypothesis that CMEP is a tumor-secreted factor. Since uterine carcinomas are associated with a poorer prognosis with plasmapheresis therapy, compared to lung and ovarian cancers, we believe that there may be heterogeneity in the CMEP factor. We hypothesize that uterine cancers may secrete a growth factor that is poorly removed by plasmapheresis, while lung and ovarian carcinomas are associated with a CMEP factor produced by the immune system (ie: IgG), which is efficiently removed by plasmapheresis.

HGF and anti-retinal autoantibodies were recently reported to be a new pathophysiologic mechanism for BDUMP. In this report, a patient with papillary renal carcinoma and BDUMP responded well to plasmapheresis with a correlative reduction in serum HGF levels and anti-retinal autoantibodies that crossreacted with HGF were identified(Niffenegger et al., 2018). However, these results are only correlations and no causative data is presented. Based on these results, we queried the HGF levels in our samples by immunoblot analysis. We found that HGF was present in pre-plasmapheresis, post-plasmapheresis, and IgG depleted fractions, and levels were nearly absent in the IgG enriched fraction (Fig 7B). Since all fractions stimulated melanocyte growth (Fig 5–6), we found no correlation between HGF levels and melanocyte growth in our patient with uterine cancer. HGF and anti-retinal autoantibodies, however, may play a role in papillary renal carcinoma, lung cancer, or ovarian cancer, and future studies are needed to further this interesting hypothesis.

In order to better understand the mechanism of SRD in BDUMP, we collected samples of subretinal fluid after surgical RD repair. We compared our SRD BDUMP samples to patients with RRD, uveitic SRD, and multiple myeloma-associated SRD. We found no enrichment of IgG heavy or light chains in SRF samples from our BDUMP patient, compared to control SRF (Fig 8). There was more IgG qualitatively detected in the BDUMP and uveitic SRD, but this is more likely related to the duration of SRD, as previously reported (Rose et al., 1990). We hypothesize that the mechanism of SRD is related to a passive transudative process, explaining the higher concentrations of albumin compared to IgG. A transudation through dysfunctional RPE corresponds to the many expanding areas of hyperfluorescence observed with UWF-FA (Fig 2–3).

(7). Conclusions

We report a case of BDUMP associated with endometrial carcinoma that responded poorly to plasmapheresis treatment. We hypothesize that disease heterogeneity among BDUMP patients exists and that uterine carcinomas may be less likely to respond to plasmapheresis because the CMEP factor is a tumor-secreted factor that is poorly removed by plasmapheresis due to fast rebound. Alternatively, lung and ovarian carcinomas have a better response to plasmapheresis, possibly because the CMEP factor is an immunoglobin removed more effectively by plasmapheresis. Finally, lack of concomitant chemotherapy is associated with a worse response to plasmapheresis because ongoing CMEP factor production will continue whether the CMEP factor is an immunoglobulin or a growth factor.

• Cultured melanocyte elongation and proliferation (CMEP) factor stimulates melanocyte growth

• CMEP, in our patient, is not concentrated in the IgG enriched fraction

• CMEP is still present in plasma after 6 plasamapheresis treatments over 9 days

• CMEP, in our patient, is unlikely to be an immunoglobulin, and is more likely a tumor-secreted growth factor

(2). Abbreviations

BDUMP

bilateral diffuse uveal melanocytic proliferation

CMEP

cultured melanocyte elongation and proliferation factor

SRD

serous retinal detachment

SRF

subretinal fluid

IgG

immunoglobulin

HMGS

human melanocyte growth serum

BCA

bicinchoninich acid

BSA

bovine serum albumin

TBS-T

tris-buffered saline – tween 20

ANOVA

analysis of variance

RPE

retinal pigment epithelium

OCT

optical coherence tomography

FA

fluorescein angiography

UWF

ultra-widefield

ICGA

indocyaninge green angiography

RD

retinal detachment

RRD

rhegmatogenous retinal detachment

SDS-PAGE

sodium dodecyl sulfate - polyacrylamide gel electropheresis

HGF

hepatocyte growth factor