MERTK Arginine-844-Cysteine in a Patient with Severe Rod–Cone Dystrophy: Loss of Mutant Protein Function in Transfected Cells

Abstract

purpose. Mutations in the MERTK gene are responsible for retinal degeneration in the Royal College of Surgeons (RCS) rat and are a cause of human autosomal recessive retinitis pigmentosa (RP). This study reports the identification and functional analysis of novel MERTK mutations to provide information regarding whether they are causative of severe rod–cone degeneration in a young patient.

methods. MERTK missense variants identified by single-strand conformational polymorphism (SSCP) and sequence analysis were introduced into expression constructs and used to transfect HEK293T cells. Recombinant protein expression was assayed with anti-MERTK and anti-phosphotyrosine antibodies. Protein turnover was assayed in pulse–chase studies of ³⁵S-methionine incorporation. Transcript levels were determined by quantitative RT-PCR.

results. Three MERTK sequence variants were identified in a patient with rod–cone dystrophy: R722X in exon 16 and R865W in exon 19 on the paternal allele and R844C in exon 19 on the maternal allele. The R844C sequence change affects an evolutionarily conserved amino acid residue and was not detected in unaffected individuals. In transfected HEK293Tcells, wild-type (wt) and W865 MERTK were expressed at equivalent levels and present in the plasma membrane, stimulated tyrosine phosphorylation, and induced significant rounding of the cell bodies. In contrast, C844 MERTK was expressed at low levels and did not stimulate tyrosine phosphorylation. In addition, the relative stability of C844 MERTK was significantly less than wt in assays of protein turnover. At age 13, the patient had 20/60 and 20/200 acuities, tunnel vision of 5° centrally, and a far temporal peripheral crescent bilaterally, and ERGs were nondetectable. The fundi showed bull’s-eye macular atrophy and widespread RPE thinning.

conclusions. The present study reports the identification of R844C, the first putative pathogenic MERTK missense mutation that results in severe retinal degeneration with childhood onset when in compound heterozygous form with a R722X allele. The loss of function of C844 MERTK is probably due to decreased protein stability.

The retinal degeneration phenotype of the Royal College of Surgeons (RCS) rat results from a 1.5- to 2-kb deletion in the Mertk gene ¹ that impairs the ability of the retinal pigment epithelium (RPE) to phagocytose membranes shed from photoreceptor outer segments (OS) during their renewal. ^{2 3 4} Shed OS membranes are bound, but not ingested by the RPE, resulting in a debris field that disrupts the vital flow of nutrients, vitamin A, and oxygen between the RPE and retina. ^{5 6} As one of the best-characterized animal models of retinal dystrophy, the RCS rat has been the subject of intense study and testing of various therapeutic strategies. ⁷ A corresponding mouse model also exists—created by targeted disruption of c-Mer (the mouse orthologue of Mertk) by using homologous recombination—that has a retinal degeneration phenotype similar to that of the RCS rat. ^{8 9}  

Mertk (Tyro12) ^{10 11} is a member of the receptor tyrosine kinase family of cell-surface receptors that includes Axl (Tyro7) and Rse (Sky, Tyro3) and consists of an intracellular kinase-containing domain, a transmembrane region, and a cell-adhesion molecule–related extracellular domain. ^{12 13 14 15} Mertk is expressed in a number of tissues outside the eye, including macrophages, epithelia, and reproductive tissue. ¹⁰ It is required for clearance of apoptotic cells by mononuclear phagocytes in mice, ¹⁶ with its absence resulting in progressive lupuslike autoimmunity. ¹⁷ It also plays a role in blocking lipopolysaccharide-induced endotoxic shock through inhibition of TNF-α production. ⁸ Growth-arrest–specific protein 6 (Gas6) is an activating ligand for Mertk and the other receptors in this family, ^{18 19} and stimulates OS phagocytosis by RPE cells in culture. ^{20 21} Gas6 may facilitate phagocytosis by binding to phosphatidylserine residues exposed on the outer leaflet of spent OS membranes, thus promoting association with the RPE. ^{1 22 23} Ligand activation of Mertk is predicted to result in the formation of receptor homodimers, activation of receptor kinase activity, and phosphorylation of tyrosine residues in the receptor intracellular domain. ²⁴ The role of Mertk in OS phagocytosis may also involve interactions with other receptors involved in OS recognition and binding, including αvβ5 integrin, type B scavenger receptor CD36, and mannose receptors. ^{25 26 27 28}  

MERTK was established as a human retinal dystrophy gene by the finding of disease-associated mutations in three patients with autosomal recessive retinitis pigmentosa (RP). ²⁹ The mutations identified were presumed loss-of-function alleles predicted to result in truncated protein lacking the intracellular region. A number of apparently nonpathogenic MERTK sequence variants were also identified, suggesting that MERTK tolerates a significant number of amino acid substitutions. The present study reports the identification of novel disease-associated MERTK mutations in a young retinal dystrophy patient. Our findings establish that the amino acid substitution found can disrupt MERTK signaling by decreasing protein stability and provide a description of the associated phenotype that may be helpful in identifying additional patients with MERTK mutations. 

Methods

Patient Collection and Evaluation

The study conformed to the tenets of the Declaration of Helsinki. Patients with retinal dystrophy and unaffected control individuals signed informed consents and were selected for participation on the basis of clinical evaluations. 

MERTK Mutation Screening

Genomic DNA was extracted from blood samples using a DNA isolation kit (Puregene; Gentra Systems, Minneapolis, MN). Primers corresponding to intronic sequences were used for PCR amplification of all 19 exons of MERTK. ²⁹ Single-strand conformational polymorphism (SSCP)/heteroduplex screening was performed with 15% glycerol–containing polyacrylamide gels at room temperature. ³⁰ DNA sequence analysis was performed by the Biomedical Research DNA Sequencing Core of the University of Michigan. 

Screens for the MERTK sequence variants c.2530C→T and c.2593C→T were performed with AciI and Hpy188I (New England Biolabs, Beverly, MA) digests of the exon 19 sequence (259 bp) that was PCR amplified with the primers (forward) 5′-ACAAAGAGATGGGTGCCATGC-3′ and (reverse) 5′-CGATGTTCAAGTCCAGTGGAG-3′. An AciI site present in wild-type (wt) DNA gave rise to 96- and 163-bp fragments and was absent in the c.2530C→T variant. Two Hpy188I sites present in wt DNA gave rise to two 49- and one 161-bp fragments, whereas one Hpy188I site was absent in the c.2593C→T variant, resulting in 49- and 210-bp fragments. 

MERTK Constructs and Expression Studies

A human cDNA encoding MERTK (including 70-bp 5′- and 273-bp 3′-untranslated sequence) was obtained by screening a kidney cDNA library and cloned into pcDNA3.1 (Invitrogen, San Diego, CA). The c.2530C→T (R844C) and c.2593C→T (R865W) sequence variants were introduced by site-directed mutagenesis (QuikChange Kit; Stratagene, La Jolla, CA). Constructs were verified by DNA sequencing, and two independent isolates of each construct were obtained and used in replicate experiments. 

HEK293T cells grown in DMEM containing 1 mM sodium pyruvate and 10% FBS were transiently transfected with MERTK constructs using a lipophilic transfection agent (1μg DNA/4 μL reagent; Lipofectamine Plus; Invitrogen) according to the manufacturer’s instructions. A lacZ encoding expression construct (pCMV-βgal) was included in each MERTK DNA sample (10% of total DNA) to control for transfection efficiency assayed by β-galactosidase staining. MERTK expression was evaluated with a rabbit polyclonal antiserum raised against a fusion protein containing the 103 carboxyl-terminal amino acids of rat Mertk. ³¹ The specificity of this antiserum for the rat protein has been demonstrated ³¹ and confirmed for the human protein by comparison of cells transfected with either empty pcDNA3.1 or pcDNA3.1 containing MERTK cDNA, by using immunohistochemical and Western analysis of MERTK expression. 

For immunohistochemical analysis, HEK293T cells grown and transfected on plastic chamber slides were incubated with a 1:100 dilution of Mertk antiserum and then with Cy3-conjugated anti-rabbit IgG (Molecular Probes, Eugene, OR). Cells were viewed and photographed with a microscope (Eclipse E800; Nikon, Melville, NY) equipped with a digital camera (DMX1200; Nikon) and the manufacturer’s data acquisition software. 

For Western analysis, transfected HEK293T cells were dissociated in 2% SDS-containing sample buffer, electrophoresed on precast Bis-Tris 10% acrylamide gels (NuPage; Novex, San Diego, CA) in the 3-(N-morpholino)propanesulfonic acid (MOPS) electrophoresis buffer system, and transferred to nitrocellulose membranes (Protran BA85; Schleicher & Schuell, Keene, NH). Equivalent protein loading was verified by Coomassie blue staining. Blots were incubated with a 1:1000 dilution of Mertk antiserum, then with alkaline phosphatase-conjugated anti-rabbit IgG (Promega, Madison, WI). Protein molecular masses were estimated by comparison of mobility relative to prestained standards (SeeBlue Plus2; Invitrogen) calibrated for the MOPS electrophoresis buffer system used. 

For analysis of the glycosylation status of recombinant MERTK, lysates from transfected HEK293T cells in 1× denaturing buffer were treated with either Endo H or PNGase F endoglycosidases, according to the manufacturer’s instructions (New England Biolabs), and Western analysis was performed using Mertk antiserum as described earlier. 

For detection of phosphotyrosine-modified proteins in MERTK-transfected HEK293T cells, Western analysis was performed on protein samples electrophoresed as described earlier, and probed using a chemiluminescence kit (Anti-pTyr Immunoblotting and ECL Detection kit; Upstate Biotechnology, Lake Placid, NY). 

For immunoprecipitation experiments, cells grown and transfected in six-well plates were lysed in 150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl (pH 8.0), and 1% NP-40, containing protease inhibitors (Complete; Roche Diagnostics, Indianapolis, IN). Lysates were cleared by low-speed centrifugation. Supernates were incubated overnight with Mertk antiserum (10 μL) and then for 2 hours with protein-A agarose beads. Bound proteins were eluted with 2% SDS sample buffer, electrophoresed as described earlier, and transferred to nitrocellulose. Blots were probed for phosphotyrosine-modified proteins, using the chemiluminescence kit (Anti-pTyr Immunoblotting and ECL Detection kit; Upstate Biotechnology) or for MERTK expression as described earlier. 

For pulse–chase analysis of protein turnover, at 16 hours after transfection, cells were incubated for 1 hour in serum-free and methionine/cysteine (met/cys)–deficient medium, then for 5 hours in deficient medium supplemented with ³⁵S-met/cys promix (Amersham, Arlington Heights, IL) and then returned to standard serum-containing medium and incubated for various times. Cells were harvested, and recombinant protein was immunoprecipitated, eluted with SDS sample buffer, electrophoresed on 10% Tris-glycine acrylamide gels, and analyzed by fluorography. 

For quantitation of MERTK transcripts, total RNA was isolated from transfected cells (RNeasy kit; Qiagen, Valencia, CA), and first-strand cDNAs were synthesized according to the manufacturer’s protocol (Amersham). PCR products (226 bp) were amplified with the following primers: forward (exon 18), 5′-CAGAACCATGAGATGTATGACTATC-3′, and reverse (exon 19), 5′-TCTCCAGCAACTGTGTATTGAC-3′. Control reactions were performed with primers for human hypoxanthine phosphoribosyltransferase (HPRT): forward (exon 4), 5′-GTGGAGATGATCTCTCAACTTTAACTG-3′, and reverse (exon 8), 5′-CATTATAGTCAAGGGCATATCCTACAAC-3′ (236-bp product). Reaction progress was monitored in real-time with a thermocycler (Rotorgene Thermocycler RG3000; Corbett Research, Mortlake, NSW, Australia) and the manufacturer’s software. 

Ophthalmic Examination

Evaluation of patient 3061 included tests of visual acuity, color vision, visual fields, dark-adapted absolute thresholds, standard clinical electroretinograms (ERGs), and fundus photography, using methods described previously. ³² Best-corrected distance visual acuity was determined with the Early Treatment Diabetic Retinopathy Study letter charts at a 4-m viewing distance. Color vision was evaluated with Ishihara plates. Central visual fields were determined with standard Humphrey 10-2 automated perimetry (Carl Zeiss Meditec, Dublin, CA). Goldmann kinetic visual fields were obtained with the targets V4e, II4e, and I4e on a standard 31.5-asb (10 cd/m²) background. Thresholds were measured at fixation and at several locations along the horizontal meridian of visual field on a dark adaptometer (Goldmann-Weekers DarkAdaptometer; Haag Streit, Köniz, Switzerland) after 45 minutes of dark adaptation. ERGs were measured according to the International Society for Clinical Electrophysiology of Vision standard, ³³ beginning after 45 minutes of dark adaptation using 10-μsec xenon flashes in a Ganzfeld bowl. Pupils were fully dilated using phenylephrine HCl (10%) and tropicamide (1%), and Burian-Allan bipolar corneal electrodes (Hansen Ophthalmic Instruments, Iowa City, IA) were applied after proparacaine HCl (0.5%) topical anesthesia. Responses were amplified at 0.1 to 1000 Hz (−3 dB) digitized and stored. Rod-predominant responses were recorded first, using 0.5-Hz dim blue stimuli (440 nm peak, 70 nm half-width, −1.86 log cd-s/m²). Photopic responses (primarily cone circuit responses) were recorded after light adaptation for at least 5 minutes with a 43 cd/m² background light in the Ganzfeld bowl and elicited with single white flashes (0.5 Hz, 1.0 log cd/m²). 

Results

MERTK Sequence Variants

In a mutation screening study of patients with autosomal recessive and sporadic forms of retinal dystrophy, three novel MERTK sequence variants were identified in heterozygous form in a single individual (patient 3061): a c.2164C→T transition in exon 16 predicting a R722X premature stop codon, a c.2530C→T transition in exon 19 predicting a R844C change, and a c.2593C→T transition in exon 19 predicting a R865W change 1 . Three common MERTK polymorphisms reported previously (N118S, R466K, I518V) ²⁹ were also detected in the patient DNA. 

Segregation analysis of the three novel MERTK sequence variants indicated that the R844C change was present on the patient’s maternal allele, and the R722X and R865W changes were present on the paternal allele. Both R844C and R865W affect residues located just downstream of the tyrosine kinase region of the receptor, ^{10 24} whereas R722X predicts a truncated protein missing more than one fourth of the full-length MERTK sequence of 999 amino acids. As R865W was present on the R722X presumed loss-of-function allele that predicts truncation of the protein before position R865, and furthermore, because R865 is not conserved in Mertk from mouse and rat (where it is glutamine), the involvement of R865W in disease pathogenesis in this patient seemed unlikely. However, R844C on the maternal allele represents a nonconservative amino acid substitution at a position that is fully conserved in Mertk from human, mouse, rat, and chicken, ^{1 10 11 34} as well as the related gene Tyro3. ^{14 15} The R844C change was not detected on 350 chromosomes from patients with retinal dystrophy or on 100 chromosomes from unaffected control individuals, whereas the R865W change was detected in heterozygous form in one patient who did not have a second apparent disease-causing MERTK mutation. The R722X and R844C sequence changes identified in the present study, as well as MERTK mutations identified in patients previously, ²⁹ are shown on a schematic of the gene structure in 1 . 

MERTK Expression Studies

Western blot analysis of lysates from HEK293T cells expressing MERTK wt or the W865 variant showed three bands of MERTK immunoreactivity (apparent molecular mass: band 1, 155 kDa; band 2, 125 kDa; and band 3, 95 kDa) that were not seen in cells transfected with empty vector (MERTK predicted molecular weight, 107 kDa; 2 ). However, cells transfected with the C844 variant showed significantly lower levels of MERTK expression, exhibiting only faint bands 1 and 2 of MERTK immunoreactivity. In all cases, bands 1 and 2 migrated with the same mobility as recombinant rat Mertk from a plasma membrane–enriched fraction from NRK49 cells transfected with an adenoviral vector-Mertk construct. ³¹ Similar results were obtained in transfected COS-7 cells, and in HEK293T cells harvested at 18 and 40 hours after transfection, but with lower overall expression levels at shorter times (data not shown). MERTK immunoreactivity was not detected in the culture medium from cells transfected with any variant, as might be expected if the recombinant protein were mistrafficked and secreted from the cells (data not shown). 

To determine whether the three bands of MERTK immunoreactivity represent differentially glycosylated forms of the receptor (14 potential sites for N-linked glycosylation are present in the human protein sequence ¹⁰ ), lysates from transfected cells were treated with either PNGase F, which cleaves nearly all forms of N-linked oligosaccharides, or Endo H, which cleaves high mannose and hybrid forms of N-linked oligosaccharides, and were examined by Western analysis 2 . PNGase F treatment collapsed the three bands of MERTK immunoreactivity into a single band migrating at the position of band 3. Endo H treatment did not affect band 1, but acted on band 2 to produce new bands migrating just above the position of band 3. These results suggest that band 3 of MERTK immunoreactivity corresponds to the unglycosylated receptor, band 1 to a mature form of the receptor modified by complex oligosaccharides, and band 2 to a (potentially intermediate) form of the receptor containing high mannose and/or hybrid oligosaccharides. These findings are consistent with previous studies showing that rat Mertk exits in two differentially glycosylated forms in plasma membrane–enriched fractions. ³¹ In the present studies, the additional finding of an unglycosylated form was probably due to analysis of total protein lysates from cells expressing high levels of MERTK that may exceed the posttranslational processing capacity of the system. The apparently lower molecular masses observed in the present study compared with those previously reported for rat Mertk (190–200 and 150–160 kDa) ³¹ reflect differences in the electrophoresis and calibration systems used. 

To determine whether MERTK expression affects protein phosphorylation in transfected cells, cell lysates were probed for phosphotyrosine-modified proteins using Western analysis and anti-pTyr antibodies. A number of intensely reactive proteins were detected in cells expressing MERTK wt or the W865 variant that were not present in cells transfected with the C844 variant or with empty vector 2 . When recombinant human MERTK was immunoprecipitated from cell lysates with anti-Mertk antiserum and subjected to Western analysis with anti-pTyr antibodies, phosphoproteins corresponding to bands 2 and 3 were detected in cells that expressed MERTK wt or the W865 variant, but not in cells with the C844 variant or empty vector 2 . These results suggest that MERTK is expressed as an active form that undergoes autophosphorylation and stimulates tyrosine phosphorylation of additional downstream signaling proteins, either directly or indirectly. This activation may be a consequence of high expression levels in transfected cells that promote receptor aggregation and dimer formation. Activation of MERTK tyrosine phosphorylation in cells expressing the C844 variant does not occur. 

Immunohistochemical analysis of cells transfected with MERTK wt or the W865 variant showed similar patterns and levels of protein expression consistent with localization of the receptor in the plasma membrane, accompanied by a dramatic rounding of the cell bodies 3 . In contrast, cells transfected with the C844 variant exhibited a low-intensity, discontinuous pattern of labeling and few rounded-up cells, with some cellular disruption apparent. Individual cells transfected with empty vector, MERTK wt, C844, or W865 are shown in 3 , and fields of cells transfected with C844 or W865 are shown at lower magnification in 3 . Analysis of transfection efficiency by β-galactosidase staining (from a cotransfected lacZ construct) showed equivalent expression levels in cells expressing MERTK wt or the W865 variant, with somewhat lower apparent levels (∼two- to threefold) consistently observed in cells expressing the C844 variant (tested using two independent clones; 3 and data not shown). Analysis of cell viability by trypan blue exclusion staining showed little cell death in cultures that expressed any MERTK variant, although it was slightly higher in cells expressing the C844 variant 3 . Similar findings were obtained at 24 and 42 hours after transfection, with lower overall expression levels at shorter times. 

The relative stability of MERTK wt, C844, and W865 proteins was determined in pulse–chase experiments involving ³⁵S-met/cys metabolic labeling of transfected cells, followed by immunoprecipitation of the radiolabeled receptor, electrophoresis, and fluorography. The half-life (t _1/2) of decay of the resultant radioactive bands was calculated by densitometry. For each of the variants, radiolabeled protein corresponding to MERTK bands 2 and 3 was immunoprecipitated at the end of the period of metabolic labeling (0 hour; 4 ). In all cases, band 3 disappeared by the first time point of chase (8 hours), most likely because of glycosylation and conversion to band 2 (that increases in abundance at 8 hours in wt). In cells expressing MERTK wt or the W865 variant, band 2 was relatively stable, with approximately half of the radiolabeled protein remaining at the end of the 32-hour chase. The apparent size of band 2 shifted downward during the chase, presumably due to phosphorylation. In cells expressing the C844 variant, the decay of band 2 was much more rapid (t _1/2 < 8 hours), with less than 10% of the radiolabeled protein remaining at the end of the chase. In contrast, transcripts encoding each of the three of the variants did not markedly differ from one another in their rate of decay (as analyzed by quantitative RT-PCR), although C844 transcripts were present in slightly lower abundance (∼twofold), consistent with the lower observed transfection efficiency of this construct 5 . 

Taken together, the results of our expression studies suggest that the MERTK R844C sequence change observed in the patient in our study results in a significant reduction in protein stability. Subsequent loss of function probably results from disruption of MERTK signaling involving tyrosine phosphorylation of cellular proteins. 

Patient Phenotype

The patient was a young girl whose parents first noticed her vision problems when she was 3 years old and began to trip over objects in front of her. When she was 8 years of age, advanced rod–cone dystrophy was diagnosed, and she had surgery to correct strabismus. At age 9 her visual acuities were 20/60 +2 OD and 20/60 +2 OS, with the myopia corrected by spectacles. She had no color discrimination. Nystagmus was present. Her lenses and media were clear and the pupils round and reactive to light. Her fundus showed considerable disc pallor from gliosis and retinal vessels constricted to approximately half normal size. Both eyes showed macular atrophy, fine bone spicule pigment in all quadrants, dense parafoveal pigmentation, and heavy RPE granularity throughout the fundus. Visual fields were considerably limited in both eyes, with a “tunnel” of 20° central detection and a peripheral crescent of detection in the far periphery of both eyes. Dark-adapted threshold sensitivity was elevated by 4 log units in both eyes. The ERG showed total loss of response to all rod and cone stimuli (>99%, to the limit of noise alone). 

By age 13, the patient’s best corrected visual acuity had decreased to 20/200, and her visual fields to the largest test target (V4e) had constricted to less than 5° central detection with a narrow peripheral crescent remaining for both eyes. Her condition qualified as a severe, early-onset rod–cone degeneration, with widespread loss of visual function at a young age. She used a white cane for mobility, attended regular school with services from the visually impaired program, and although intelligent, struggled with her studies as she continued to rely on vision for reading rather than learning Braille. At age 15, both of her fundi showed a pattern of bull’s-eye macular atrophy with widespread RPE thinning in the periphery 6 . Her medical history included kidney reflux diagnosed at 13 months of age and managed with antibiotics. Both parents, one sibling 2 years older, and two half-siblings 4 years younger, had normal vision. 

Discussion

The identification of the RCS rat disease gene and the finding of MERTK mutations in patients with RP established the relevance of defects in RPE phagocytosis in human retinal disease. ^{1 29} The MERTK mutations first identified—a 5-bp deletion, a splice-site defect, and a premature stop codon—were predicted to result in loss of function due to truncation or absence of the protein. The present study reports the finding of a most likely missense mutation, R844C, present in compound heterozygous form with a nonsense mutation, R722X, in a young patient with severe rod–cone dystrophy. The R844C change results in decreased stability of the recombinant protein in transfected cells. Our data therefore suggest that loss of function in vivo is likely due to loss of MERTK signaling. 

Our studies of MERTK sequence variants in transfected cells showed that overexpression of the protein dramatically increased tyrosine phosphorylation of cellular proteins, including the receptor itself. Activation of MERTK tyrosine kinase activity may have resulted from high levels of expression leading to aggregation and receptor dimer formation (similar to that resulting from antibody induced receptor aggregation ¹⁹ ), and subsequent autophosphorylation and downstream signaling. This interpretation is consistent with the finding that an unglycosylated form of the receptor is phosphorylated that is unlikely to be present on the cell surface. It is also possible that an activating ligand was present in the HEK293T and COS-7 cultures that further stimulated MERTK tyrosine kinase activity. This activating ligand may have been Gas6, the known MERTK ligand or a structurally related member of the same family that includes protein S. ³⁵ The presence of significant amounts of activating ligand may be unique to cultures of SV-40–transformed cells, since high levels of tyrosine kinase activity were not observed in previous studies of recombinant MERTK in HEK293 and CHO cells before addition of exogenous ligand. ¹⁹  

Cells expressing MERTK exhibited a dramatically rounded shape, but remained viable for at least 2 days after transfection. Others have also noted that MERTK signaling results in morphologic changes in a hematopoietic cell line. ³⁶ The shape changes we observed probably resulted from MERTK chronic stimulation of tyrosine phosphorylation. In contrast, cultures transfected with the C844 variant exhibited fewer rounded cells and no increase in tyrosine kinase activity. However, many C844 variant-expressing cells appeared disrupted and exhibited aberrant shapes. One possible explanation is that overexpression of inactive protein compromised cellular integrity or created an unsustainable metabolic load. Another possibility is that overexpression of the C844 variant resulted in inactivation of endogenous MERTK necessary for normal cell function, perhaps by forming nonfunctional dimers. In this regard, it will be interesting in future studies to determine whether autosomal dominantly inherited MERTK-associated disease exists. 

The present findings highlight the critical role of tyrosine kinase activity in the mechanism of RPE phagocytosis and point to similarities with the mechanism of uptake of apoptotic thymocytes by macrophages. ¹⁶ In macrophage phagocytosis, a role for MERTK signaling could be to regulate actin assembly and membrane remodeling necessary for pseudopod extension and phagosome closure. On the molecular level, activation of MERTK tyrosine kinase activity in transfected hematopoietic cells has been shown to result in Grb2 recruitment and activation of PI 3-kinase, NF-κB, and mitogen-activated protein (MAP) kinase. ²⁴ It is also likely to involve additional signaling molecules, including small G-proteins and intracellular tyrosine kinases. ³⁷ Future studies are needed to elucidate the MERTK signaling pathway relevant to RPE phagocytosis and to characterize those aspects potentially unique to the retina and critical for its survival. As the corresponding RPE proteins are identified, each will provide a new candidate disease gene for consideration. Cell culture assay systems, such as that reported in the present study, should prove useful for evaluation of additional signaling components and corresponding mutations. 

Patients with mutations in MERTK exhibit profound loss of both rod and cone function. Such individuals, as well as, potentially, those with genetic defects at other points in the phagocytic mechanism, may be candidates for treatment options for which there is first evidence of effectiveness in RCS rat, including subretinal transplantation of RPE cells and cell lines, ^{38 39 40} treatment with various growth factors (most notably bFGF and CFTR), ^{41 42 43 44} and adenovirus–mediated gene replacement therapy. ⁴⁵ The latter approach is on the fast track with another RPE disease caused by mutations in the RPE65 gene, ⁴⁶ based on preliminary successes in the naturally occurring mutant dog that show partial recovery of vision after subretinal injection of Rpe65-encoding, adeno-associated viral vector constructs. ^{47 48} There is high optimism that clinical trials of RPE65 replacement therapy will begin in the near future in a first group of patients, as improved quality of life may be achieved by even partial rescue of a limited region of the retina. The identification of additional retinal dystrophy patients with mutations in MERTK, or other downstream genes, may establish another cohort of individuals whose disease may be treatable by similar approaches. The expression of MERTK outside the visual system and the possibility of functional studies in tissues of patients (e.g., monocytes) may offer important advantages in implementing and validating successful, relevant therapeutic protocols. 

 Supported by National Eye Institute Grants R01-EY12298 (DAT) and R03-EY014650 (DV), Vision Research Core Grant P30-EY07003, and Grant M01-RR00042 to the University of Michigan, General Clinical Research Center; and by the British Retinitis Pigmentosa Society (AG, DAT), Deutsche Forschungsgemeinschaft (AG), Foundation Fighting Blindness (AG, PAS, DAT, DV), Research to Prevent Blindness (DAT), and the Karl Kirchgessner Foundation (DV).

 Submitted for publication August 20, 2003; revised December 24, 2003; accepted January 22, 2004.

 Disclosure: C.L. McHenry, None; Y. Liu, None; W. Feng, None; A.R. Nair, None; K.L. Feathers, None; X. Ding, None; A. Gal, None; D. Vollrath, None; P.A. Sieving, None; D.A. Thompson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Figure 1.

 

MERTK sequence variants identified in a patient with severe rod and cone degeneration. (A) DNA sequence chromatograms of the patient (top) and an unaffected control (bottom). (B) Schematic of MERTK gene structure showing the mutations identified in the present study (bold) and in a previous study.²⁹ FN, fibronectin; TM, transmembrane; TK, tyrosine kinase.

Figure 2.

 

Analysis of recombinant MERTK expression and phosphorylation. HEK293T cells were transfected with pcDNA3.1 constructs encoding human MERTK wt, C844, or W865 protein, or with empty vector. (A) Western blot analysis of cell lysates using Mertk antiserum. Human MERTK proteins are shown alongside rat Mertk from NRK49 cells transfected with an adenoviral vector construct and prestained molecular mass standards. (B) Western blot analysis of lysates from cells transfected with wt MERTK, treated with the endoglycosidase EndoH or PNGase F, then probed with Mertk antiserum. (C) Western blot analysis of cell lysates from (A) using anti-pTyr antibodies that detect phosphotyrosine-modified proteins. (D) Western blot analysis of proteins immunoprecipitated from cell lysates using Mertk antiserum, probed using anti-pTyr antibodies. Results shown are representative of those in five independent analyses at 40 hours after transfection.

Figure 3.

 

Immunohistochemical analysis of recombinant MERTK expression. HEK293T cells were grown on plastic chamber slides and transfected with expression constructs encoding MERTK wt, C844, or W865 protein or with vector without insert. (A, B) Fixed cells incubated with Mertk antiserum and then with Cy3-conjugated anti-rabbit IgG. (C) Fixed cells stained for β-galactosidase activity resulting from cotransfected pCMV-βgal. (D) Living cells tested for trypan blue exclusion. Cells were viewed and photographed on a fluorescence-light microscope equipped with digital camera and data-acquisition software. Results shown are representative of those in four independent analyses.

Figure 4.

 

Relative stability of recombinant MERTK in pulse–chase experiments involving ³⁵S-met/cys metabolic labeling of transfected HEK293T cells. MERTK was immunoprecipitated from cells after 0, 8, 16, 24, and 36 hours of chase and analyzed by fluorography after electrophoresis on SDS-polyacrylamide gels. Results are representative of those in four independent analyses. Arrow: position of band 2.

Figure 5.

 

Relative abundance of MERTK transcripts in transfected HEK293T cells. RNA was isolated from transfected cells harvested during the pulse–chase regimen and quantitated by real-time RT-PCR analysis. Each time point was sampled and analyzed in triplicate (nine data points) and normalized to HPRT. Threshold cycle values are relative to samples with vector without insert. The data are representative of three independent analyses with standard deviations shown. MERTK wt (•); W865 (▪), and the averaged results of two independent clones of C844 (▴).

Figure 6.

 

Fundus photographs (A, right and B, left eyes, respectively) of a patient with R722X and R844C mutations in MERTK showing bull’s-eye macular atrophy and widespread RPE thinning. (C) Visual fields for the patient’s right eye exhibited progressive constriction of both central and peripheral islands of vision between 9 and 13 years of age, tested with the large V4e target on a Goldmann perimeter. She had scant central perception of the small I4e target at age 9 but no longer at age 13. Field perception for the left eye was comparable.

 

The authors thank the patients and their families for participating in this study and Misao Higashi for technical assistance with DNA sequencing.