Phenotype in Retinol Deficiency Due to a Hereditary ...

Phenotype in Retinol Deficiency Due to a Hereditary Defect in Retinol Binding Protein Synthesis Mathias W. Seeliger,x Hans K. Biesalski,2 Bernd Wissinger,1 Harald Gollnick? Stephan Gielen,4 Jiirgen Frank,1 Susanne Beck,^ and Eberhart Zrenner1 PURPOSE. TO describe the phenotype caused by a retinol deficiency in a family with compound heterozygous missense mutations (Hc4lAsn and Gly75Asp) in the gene for serum retinol binding protein (RBP). METHODS. The two affected sisters. 17 (BR) and 13 (MR) years old, were examined clinically and with perimetry. color vision tests, dark adaptometry, rod- and cone-isolated electroretinograms (ERGs), multifocal ERGs, electrooculograms (EOGs), and laboratory tests. RESULTS. There were no complaints besides night vision problems and no history of systemic disease. Visual acuity was reduced to 20/40 (BR) and 20/25 (MR). Anterior segments were normal except for a discrete iris coloboma. Both patients showed a typical "funclus xerophthalmicus." featuring a progressed atrophy of the retinal pigment epithelium. Dark adaptation thresholds were elevated. In the scotopic ERG, only reduced mixed responses were recordable. The photopic ERG was reduced in BR and normal in MR; implicit times were highly (BR) to slightly (MR) elevated. There was no (BR) to little (MR) light reaction in the EOG. Ati-trans retinol levels were 0.19 /AM and 0.18 ju,M (normal range, 0.7-1.5 /LIM) for BR and MR, respectively, and did not increase in a dose-response test. RBP was below detection threshold, and retinvl esters were normal. CONCLUSIONS. Both affected siblings had no detectable serum RBP, one sixth of normal retinol levels, and normal retinvl esters. The retinal pigment epithelium was severely affected, but: besides acne there were no changes to other organs. This gives evidence for an alternative tissue source of vitamin A, presumably retinvl esters from cliylomicron remnants. The normal retinol levels in the tear fluid explain the lack of xerophthalmia. However, considering the role of RBP in the tear fluid and, during development, in the yolk sac there is also evidence that there are organ-specific RBP forms not affected by the genetic defect. (Invest Ophthalmol Vis Sci. 1999;4():3-1 I)

T

he effects of a nutritional vitamin A deficiency have been known for a long time. The major symptom, night blindness, was already recognized in ancient Egypt around 1500 BC.1 Another typical finding, conjunctiva! xerosis. was described in the 19th century by Hubbenet and Bitot1 and named after the latter, Bitot's spots. The characteristic changes of the posterior segment were reported in 1928.2 It was found that all these conditions respond well to liver or derived products, which contain, because of its storage function, large

From the 'University live Hospital. Department II. Tubingen. Germany; the 2IX'part.meni of Biological Chemistry and Nutritional Science. Hohenheim University, Stuttgart. Germany; the -'Department of Dermatology. Otto-von-Gucrickc University. Magdeburg. Germany; the''Department of Internal Medicine. Ruprccht-Karls-Ijnivcrsity. Heidelberg, Germany; and the sMax Planck Institute for Biology, Tubingen, Germany. Supported by Deutsche Forschungsgemeinschaft. Bonn. Grants Ru 457/1-3, Zr 1/7-3. and Zr 1/10-1. Part of this work was presented at the Eric K. Fcrnstrom Symposium "Tapetoretinal Degenerations," Lund. Sweden, June 1997. Submitted for publication May 29, 1998; revised August 7, 1998; accepted August 27, 1998. Proprietary interest category: N. Reprint requests: Mathias Secliger. University Eye Hospital, Department II, Schlcichsinisse 12-16, D-72076 Tubingen, Germany. limrsiigativc Ophthalmology & Visual Science. January 19°9. Vol. -10. No. 1 Copyright © Association for Ut'search in Vision and Ophthalmology

amounts of vitamin A. Besides its importance for the visual cycle, vitamin A has a number of important functions throughout the body. These include growtli and differentiation of epithelial tissues and bone, reproduction, and embryonic development.^'4 Ingested vitamin A enters the circulation mainly as long-chain fatty acid esters (retinvl esters) in the cliylomicron fraction. Most of the retinol is then stored in the liver in esterified form. The release from the hepatocytes requires the association of retinol (derived by hydrolysis of the esters) with retinol binding protein (RBP) synthesized also by the liver. This complex, known as holo-RBP, binds subsequently to tninsthyretin (TTR). TTR is a 55-kDa protein with two high-affinity binding sites, one for thyroxin and one for holo-RBP, that is believed to protect bound substances from metabolism and renal excretion. 5 Many organs, such as liver, kidney, small intestine, lung, spleen, eye, and testes. were shown to have specific binding sites for holo-RBP on the cell surface that promote permeation of retinol into the cell and binding to cyctoplasmic: RBP. However, brain, muscle, thymus, fat, and heart do not contain cyctoplasmic RBP (see Refs. I and 3 for an overview of the retinol metabolism). Any condition that interferes with ingestion, absorption, storage, or transport of vitamin A can lead to a deficiency in target tissues. In the eye, anterior and posterior segments both

4

Seeliger et al.

IOVS, January 1999, Vol. 40, No. 1

appear to be affected early. In the posterior segment, unavailability of retinol causes night blindness because the steady loss of rhodopsin, of which vitamin A is an essential component, during the rod visual cycle cannot be sufficiently replaced. The impact of this mechanism on the electroretinogram (ERG) has been demonstrated in animal experiments 6 and in different conditions of nutritional vitamin A deficiency in humans. /M A severe, prolonged deficiency results in degeneration of retinal pigment epithelium (RPE) and photoreceptors. The nature of anterior segment changes is not fully understood despite the fact that they are not less common. Conjunctiva! and corneal keratinization and the loss of goblet cells reflect the changes in epithelial architecture throughout the body. The pathogenesis of keratomalacia, which features corneal ulceration and necrosis, is still unclear.' In the present report, the phenotype of two sisters with a compound heterozygous mutation in the gene for serum RBI' is described. Although retinol transport is severely impaired, most tissues including the anterior segment appear mildly or not affected, whereas there is a severe degeneration of the RPE in the posterior segment.

SUBJECTS AND METHODS

Subjects Two sisters, 17 (BR) and 13 (MR) years old, were referred to our clinic by a local ophthalmologist because of unclear fundus changes. After a thorough examination, it was concluded that these sisters did not have a common hereditary retinal degeneration, and thus a search for possible causes was started. During the course of this search, the complete absence of RBP and very low levels of retinol were found in both siblings. Thus, we suspected a hereditary disorder in the retinol-RBP pathway, which was then supported by the molecular genetic investigation of the R.BP4 gene. When the nature of the disorder became clear, informed consent was obtained from all participating family members after the character and possible consequences of the investigations were explained to them. The research followed the tenets of the Declaration of Helsinki.

minutes at an illumination of 30 cd/m 2 . Electrooculogram (EOG) recordings of the slow oscillations were also performed following the respective ISCEV standard. 12 Multifocal ERG responses were obtained with the VERIS system (EDI, Inc., San Mateo, CA) implementing a multi-input stimulation technique. l3 The setup has been described in detail elsewhere. 14 Essentially, the stimulating 6\ hexagonal elements were presented on a Sony 20-inch monitor (75-Hz frame rate, 100 cd/m 2 for white, mean luminance 5 1 8 cd/m 2 ), projecting to a visual field area of 30°. Recording at 8 samples per frame yielded a temporal resolution of 1.67 msec. Laboratory tests of blood samples included an initial evaluation of routine parameters and a serologic screening. With the use of high-performance liquid chromatography (HPLC), the vitamin A status was determined twice within a month in an external laboratory. Because retinol levels were below detection threshold (0.2 jiM) both times, the following measurements were performed in the laboratory of the Department of Biological Chemistry and Nutritional Science of Hohenheim University (Stuttgart, Germany) with a 10-fold more sensitive setup (detection threshold, 0.02 /xM). RBP measurements were performed in both laboratories with a radial immunodiffusion kit (detection threshold, 0.5 JU.M; Behring Diagnostics, Marburg, Germany). Rctinyl esters were determined in the laboratory at Hohenheim University only, again using HPLC. Further details of the laboratory analysis can be found elsewhere. 15 The general clinical examination included a sonography of the abdomen, an echocardiography, and functional lung tests. Dermatological status was obtained also. The RBP4 gene encoding the serum RBP has been mapped to human chromosome 10q24. Linkage analysis was performed with close markers D10S571 and D10S185 including two unaffected siblings and the healthy mother. Direct analysis of the RUP4 gene was performed by complete sequence analysis of the coding exons and flanking intron sequences. Parts of the RBP4 gene were amplified from total DNA and directly used for sequencing. Further details of the genetic analysis can be found elsewhere. 15

RESULTS

Patient History

Methods 9

For each patient, an extended questionnaire for the collection of medical and eye histories, covering aspects such as night blindness, visual field loss, glare sensitivity, color vision, was completed. Also, the results of a full eye examination and psychophysical tests including visual acuity, color vision test (Farnsworlh D15). and visual field tests (Goldmann and Tubingen automated perimeter) were obtained. Dark adaptation was tested using a yellow target (585 nm, 500 msec duration, 2.1° in diameter) presented at 20° in the nasal field of view after 3 minutes' exposure to a white Ganzfeld of 634 candela per square meter (cd/m 2 ). Thresholds and standard errors, corresponding to a frequency of seeing of 50%, were determined by an adaptive '"up-down-stairca.se" procedure and subsequent estimation of thresholds with the maximum likelihood method. 10 Ganzfcld ERG responses were obtained following the International Society for Clinical Elect rophysiology of Vision (ISCEV) standard protocol. 1 ' Specifically, patients underwent a dark adaptation of 30 minutes before scotopic recordings, and photopic recordings were preceded by a light adaptation of 10

The two girls, 17 (BR) and 13 (MR) years old, both had a history of night vision problems since early childhood. The younger patient (MR) underwent hernia surgery at the age of 4 weeks. There were no other complaints and no history of systemic disease. No other members of the family were known to be affected.

Morphology Anterior segments were normal except for an inferior coloboma of the iris. Most prominent in the right eye of MR (Fig. 1A), it was also present to a lesser extent in the left eyes of both girls. There were no signs of xeroplithalmia, keratomalacia, or Bitots spots. In the posterior segment, the optic discs were vital but somewhat prominent and had capillary changes. There were irregular macular reflexes and no foveolar reflexes, and the whole fundus showed a patchy atrophy of the RPE (Figs. IB, 1C). In the periphery, small white clots, representing a more focal loss of RPE, were found (Fig. ID). The general clinical examination did not reveal any pathologic changes. In the sonography of the abdomen, macroscopically


Hereditary RBP-Induced Retlnol Deficiency

5

FIGURE 1. Morphologic findings. (A) Anterior segment of the right eye of patient MR, showing an iris coloboma. No signs of xerophthalmia are visible. Right fiindus of MR (B) and left fiindus of BR (C), depicting vital but somewhat prominent optic discs, irregular macular reflexes, and a patchy atrophy of the retinal pigment epithelium. (D) Detail of the right fiindus periphery of MR. The small dots represent the focal loss of retinal pigment epithelium, (E) Sonographic picture of liver (upper left) and right kidney (upper right) of MR. No macroscopic changes are visible. (F) Central forehead of MR, showing the severe comedonic acne (see text for details).

visible abnormalities of the liver (Fig. IE), kidneys, spleen, and bladder could be ruled out. There were no changes to heart chambers, valves, or related vessels and no septum defects in echocardiography. Pulmonary function was assessed by body plethysmography, spirometry, and blood gas analysis, all of which gave normal results. The dermatological examination revealed a severe comedonic acne with densely aggregated open and closed comedones, mostly of the sandpaper type, which was more intense in the younger sister. The highest

degree of change was seen in the face (Fig. IF), followed by the shoulders, back, and chest where the least changes were found. Both patients also had a widespread follicular keratosis at the extensor surfaces of the arms, on the lower back, the flank, and upper buttocks.

Psychophysics Visual acuity was moderately reduced, to 20/40 (BR) and 20/25 (MR). Visual fields revealed no defects in Goldmann perinietry

6

10VS, January 1999, Vol. 40, No. 1

Sccligcr et al.

f\

D

DUEOII

D

m E ei • •

D

ao - H •5 "2 ~ •5 c

-3 —

—

-f, •

' 10

20

30

Time of dark adaptation |minutes|

G

0 p

I ' I ' I ' 10 20 30 Time of dark adaptation [minutes]

I -SO

H

FIGURE 2. Psychophysical examination. Static perimetry of the central 30° in 13R (A) and MR (B; see text for details). Goldmann perimetry (stimulus IIl4e) in 13R (C) and MR (D). Dark adaptation curve for 13R (E) and MR (F). Standard deviations are shown at each time of measurement. I'or comparison, the 5% to 95% percentile range ol the normal control group is shown as shaded area and the median as a black line. After an initial adaptation ending approximately 1 log unit above cone threshold, there is no further improvement over time. Tritan deficiency in 13R (G) and MR (H) r revealed by the saturated panel D15 color vision test. All data shown are from left eyes.

using the IIKle stimulus (Figs. 2A, 2B). Static perimetry of the central 30° was almost normal in MR, only a slight increase of increment threshold of less than 5 dB was observed (Fig. 2C).

The extent of effect was more pronounced in BR, who showed relative defects of approximately 10 clB in the central 5° to 7°. There were additional defects in the upper and lower halves of

IOVS, January 1999, Vol. 40, No. I the 30° field outside approximately 15° eccentricity (Fig. 2D). In patient MR, dark adaptation ceased after 4 minutes, at approximately I log unit above normal cone threshold, without any change during the next 30 minutes (Fig. 2E). In patient BR, the same final threshold was reached after 10 minutes (Fig. 2F). Color vision, analyzed with the panel D15 tests, showed a tritan deficit in both patients (Figs. 2G, 2H).

Electrophysiology The scotopic ERG was not detectable using a standard flash attenuated by 24 dB. With a full standard flash, it was reduced to 20% of the normal median in MR and
Laboratory Investigation Rctinol and RBP levels, tested twice in an external laboratory by HPI.C and immunodiffusion, were below detection thresholds (0.2 JLLM and 1.1 mg/dl, respectively). With a limit of detection of 0.02 /xM, a third measurement yielded iil\-trans rctinol levels of 0.19 juM (BR) and 0.18 /u,M (MR; normal range, 0.7-1.5 ju.M). Fasting retinyl esters were 0.14 /aM (BR) and 0.22 ju.M (MR; normal range 0.05-0.25 juM). RBP was below detection threshold again. In the mother, retinol levels were 1.38 ju.M, retinyl esters 0.16 ju.M, and RBP 2.2 mg/dl (normal range, 3.0-6.0 mg/dl). A close-response test with oral application of 30,000 IU retinyl palmitate led only to a minor increase in retinol. 15 All other laboratory and serologic results were normal. In the tear fluid, retinol was present at normal levels (0.1

Genetic Analysis Segregation analysis supported linkage with the disease phenotype in the family under the assumption of autosomal recessive inheritance (maximum logarithm of the odds score 0.8). Both affected siblings showed an identical allele composition for both markers, whereas the unaffected siblings inherited different haplotypes. In accordance with this result, direct analysis of the RBP4 gene revealed the presence of two compound heterozygous missense mutations, Ile4lAsn and GIy75Asp, in both affected siblings.15 The Ilc4lAsn mutation was also found in the mother and the unaffected sister, whereas no mutations were present in the unaffected brother. Both mutations concerned amino acid positions conserved in evolution and were absent from 100 healthy control subjects (200 chromosomes).

DISCUSSION The phenotype of both affected siblings features an impairment of the posterior eye segment, iris colobomata, and acne

Hereditary RBP-Induced Rctinol Deficiency

7

but no manifest involvement of other organs. The isolated retinol deficiency found in these patients, on the basis of a lack of RBP, appears to be the cause of these changes. Morphology In the posterior segment, the syndrome includes prominent, vital optic discs with irregular capillaries. The same disc appearance was described in a 25-year-old Indonesian woman suffering from nutrition-induced panocular xerophthalmia and has been attributed to previous vitamin A therapy. 17 However, the two patients did not receive such therapy before the time of ophthalmologic examination. The most obvious finding was the atrophy of the RPE, which appeared patchy at the posterior pole and more focal in the periphery. The small yellowishwhite clotlike lesions were morphologically identical to those described in individuals with a severe nutritional vitamin A deficiency. 1 ' 1 " In the anterior segment there were no apparent changes to conjunctivae, cornea, lens, or anterior chamber. However, the presence of iris colobomata suggests that there were vitamin A deficits in week 7 to 8 of fetal life already. Studies on developmental malformations in rats induced by maternal vitamin A deficiency have shown that the eye is the structure most frequently involved (in mild cases the only one).' Also, diaphragmatic hernias were found in one third of cases. Furthermore, the lack of yolk sac RBP has been demonstrated to have a strong impact on general and eye morphologies in mice. 19 Given that the yolk sac plays a comparable role in humans, it seems unlikely that this form of RBP was missing during development. Presumably, the colobomata developed after the hemochorial placenta had functionally replaced the yolk sac placenta, so that the lack of serum RBP became functionally important. The slight degree of the changes could then be explained by a potential amelioration of the vitamin A deficit by maternal supply via the hemochorial placenta. The dermatological findings were not typical for a vitamin A deficiency. Keratosis follicularis can often be seen in individuals with atopic constitution. However, the girls' and the family's histories revealed no evidence in this regard. The widespread distribution of keratotic follicles reminds one of vitamin A-deficiency signs in phrynoderma, but in a more discrete clinical expression. The most obvious finding was the severe comedonic acne. Both clinical findings could give evidence of a disturbed keratinization process that will be further investigated. Function The electrophysiological examination revealed that the EOG, an indicator of RPE function in the presence of working photoreceptors, 20 was much more reduced than the (photopic) ERG. This finding is in accordance with the morphologic impression that the RPE might be most and probably primarily affected. Because retinol uptake and transport to the photoreceptors were severely impaired, there was no sign of rod function in dark adaptation or scotopic ERG. Because of the persistence of the disorder, a presumably secondary retinal degeneration has developed in both patients, which appears to be more severe in the older sister. The moderately reduced visual acuity, the tritan deficit, the impairment of visual fields, and the reduction of amplitudes and latency prolongation in the ERG can be interpreted as signs of such a degeneration. The poor visual acuity in patients with similar fund us changes

8

Seeliger et al.


.... i

. i.... i.... i.

i.... i 5 ms/div

5 ms/div

Ol's

ORs

2 t

»•

• - - -

r\

• • ! . .

•

• .

' .

• • : ..i : . i !

2 ms/div

IWl—J^J

i

D

I:±.^_ 2 ms/di\

\l\\W

0

D

'.29r:'.iec

ei-.iriciiy group |-|

KAY • :

OKAY

a fC

:

: . •::

: : ' : ' : ' : '

G

B

OS

ro

•

H

FiGUKi- 3. lilectrophysiological tests. Standard I-RG results: (A) UK {left column) and (B) MR (>7^»/;/ column). Top row: scotopic responses to a standard flash attenuated by the (actor indicated. No pure rod response is observable. Bottom row: oscillatory potentials (OFs) and photopic standard llasli response. The photopic response of MR is within the normal range, whereas it is diminished and substantially delayed for MR. (See text for details.) Mullifbcal electroretinogiam (HRG): Gone functional topography within a 30° visual field in MR (C) and MR (D). Each waveform represents an liRG signal evoked from the corresponding retinal area (see Ref. 14 for details on method). (E, F) Foveal response (I) and averaged responses (2-5) from the four eccentricity groups (concentric rings around the fovea). For comparison.

70KS", January 1999, Vol. 40, No. 1

described in the literature was, in contrast, probably primarily due to anterior segment problems not encountered in this pedigree. The topography of multifocal ERG responses yielded different results in both patients that might, because of the age difference, reflect two different stages of the same disorder. In MR, pericentral regions were affected most, leading to a decrease in amplitude and a slight increase in implicit time. In BR, there was a more pronounced decrease of amplitudes so that only a few peripheral responses were discernible from noise. The averaged data revealed that implicit times were severely delayed to an extent that comes close to those seen in retinitis pigmentosa.|/fK>

Hereditary RBP-Induced Retinol Deficiency

9

cannot be the sole source of retinoid. The third possibility, the use of dietary (chylomicron) retinyl ester by peripheral tissues, appears to be the most probable alternative.

Ophthalmic Pathophysiology From an ophthalmologist's point of view, the most striking question is why anterior and posterior segments are so differently affected. There are several possible explanations for the selective impairment of RPE and retina. For example, the metabolic detour may not be as effective as the normal pathway. Differences between tissues would be related to dependence on retinol. Because the RPE has a higher demand than almost all other tissues, it would be heavily affected. Or, because there are no circulating RBP-retinol-TTR complexes, highly specialVitamin A Status ized transport mechanisms are ineffective. Tissues dependent on specific receptors like the RPE2526 would be affected most. In summary, both affected siblings had no measurable serum This is supported by experiments showing that exogenously RBP, one sixth of normal retinol levels, and normal retinyl supplied retinyl esters cannot be taken up by cultured RPE esters. Genetically, the defect has been located in the RBP4 cells.25 Also, RA given as a substitute for retinol in the vitamin gene. The reduction of RBP in the mother to almost exactly A-deficient rat does not prevent retinal and RPE degeneration, one half of the mean of the normal range is in good accordance whereas other tissues (with the exception of reproductive with her heterozygous state. Because both patients are comtissue) do accept RA.r> Finally, in the strain of TTR knockout pound heterozygotes, they do not have a functioning copy of mice mentioned above, there were no eye abnormalities (even the RBP4 gene, which means that there is no carrier of retinol in histology) despite similar plasma levels of retinol and RBP. from the liver to peripheral tissues. Consequently, the doseGiven that in the absence of TTR there is at least some release response test resulted in a normal rise of retinyl esters as an of RBP-retinol complexes from the endoplasmic reticulum of indication of normal absorption, but no such rise of retinol.13 the liver, each complex would perform at least one pass Obviously, this condition is not lethal as suspected in a report 21 the circulator)' system even if TTR is needed to prethrough of a heterozygous patient who developed keratomalacia after vent glomerular filtration or metabolic degradation in the kidmeasles. The rather mild character of changes caused by such neys. In contrast to our patients, this would possibly provide a low retinol levels, at least up to the age of 17 years, is remarksmall amount of RBP-bound retinol to receptor-dependent tisable. Interestingly, results very similar to the ones presented 222 sues like the RPE. The 20% reduced retinoid content in the eye here were obtained in TTR knockout mice. * These mice cups of the TTR knockout mice could be the result of such a had plasma retinol levels of 6% and RBP levels of 5% of those mechanism. A more speculative hypothesis is based on the fact found in wild-type animals. In contrast to our results, plasma ii\\-t.rans retinoic acid (RA) increased by 2.3-fold. To the sur- that, although the interphotorecepior retinoid-binding protein is thought to accomplish most of the retinoid transport in the prise of these investigators, they could not detect any phenointerphotoreceptor matrix, there is at least a preferential apical typic abnormalities, and these mice were fertile and reprorelease of serum RBP and TTR by the RPE cells in vitro.27 The duced normally. When put on a vitamin A-deficient diet, they function of these proteins in the retina is not known, however; developed symptoms at the same time as control animals and the genetic defect might also impair the secretion of RBP by also similarly recovered after retinoid substitution. Furtherthe RPE and eventually cause retinal or RPE damage, or both. more, tissue levels of retinol and retinyl esters were normal in Other investigators have confirmed that the RBP and 1TR all tissues except eye cups, in which a 20% reduction was secreted by the RPE are identical in structure to the serum found. RBP was elevated by 60% in the liver, whereas levels in proteins.28 the kidney were normal. The authors propose three mechanisms of how these mice might manage to take up dietary The normal retinol levels in the tear fluid explain why retinoid and transport retinol to their tissues at rates sufficient the anterior eye is not affected. Previous work in rabbits has to maintain normal vitamin A status: First, tissue uptake could demonstrated that the retinol in tear fluid originates from be 2.5-fold more efficient because of a faster dissociation of the Iacrimal gland in which vitamin A is stored in the form retinol from RBP in the absence of TTR. This is not very likely of fatty acyl esters of retinol.29 The mechanism of uptake to contribute too much because the patients presented supfrom the blood is not fully understood, but because there is posedly have a similar tissue supply in the presence of TTR. no cellular RA-binding protein in the rabbit,M) it is unlikely Second, nil-trans RA might compensate for the relative absence that RA plays a role in the vitamin A supply via the tears. In of retinol in the circulation. However, because the mice reprothe absence of serum RBP, an uptake of retinyl esters from duce normally2'' and do not show any eye symptoms,6 RA the blood appears to be the most plausible mechanism. In

the normal range is shown as 5% to 95% interval (shaded area) and its median (black line), lulled symbols refer to left eyes and open symbols to right eyes. Results from MR are represented by circles and those from MR by squares. Missing values indicate that a clear peak was not present. (E) Response density plot. All values are subnormal, and the central responses in BR are missing. (F) Corresponding peak implicit times. There is a general delay in BR but only a regional delay in MR (groups 2-4). (See text for details.) Elecirooculognmi shows slow oscillations in BR (G) and MR (H). There was no (BR) to little (MR) light reaction. The Arden ratio was 0.9 in BR and 1.2 in MR (normal range, 2.1-3.0).

10

Seeliger et al.

contrast, patients with nutritional vitamin A deficits have neither retinol nor retinyl esters and therefore develop anterior xerophthalmia. However, because the vitamin A metabolism of the lacrimal gland differs somewhat between species,31 it is nor known to which extent animal model findings apply also to humans. It is remarkable that in the rabbit, retinol is secreted into the tears bound to RBP synthesized in the lacrimal gland itself.32 In contrast, the investigators were unable to detect any production of TTR. It was suggested that TTR would not be necessary because there is no glomerular filtration of tears, but this would also be true for the extracellular matrix into which RBP and TTR are secreted by the RPE.27 Furthermore, the analysis of the distribution of intraocular RBP and TTR has revealed a strict colocalization in the rat, providing strong circumstantial evidence that RBP and TTR may function cooperatively in the intraocular translocation of retinol.33 The only exception was found in the cornea, in which TTR immunoreactivity was restricted to the endothelium, whereas intense immunostaining of RBP was present in endothelium and epithelium.33 However, the epithelial staining was dependent on the method of tissue preparation, so that the investigators suspect that this RBP may be present in a different physicochemical state from RBP elsewhere in the eye. In summary, there is some evidence that the RBP in the tear fluid is different from the other RBP forms, and it can be hypothesized that it might therefore not be affected by the genetic defect.

CONCLUSIONS

IOVS, January 1999, Vol. 40, No. 1 A mouse model would thus certainly be useful to investigate the different approaches. References 1. Summer A. Nutritional blindness: xerophthalmia and kcratomalacia. In: W Tasman, EA Jaeger, eds. Duanes's Clinical Ophthalmology. Revised ed. Philadelphia: Lippincott-Raven Publishers; 1996: 1-10. 2. Uycnuira M. Uebcr eine merkwiirdige Augenhintcrgrundsveranderiing bei zwei Fallen von idiopathischcr Hcmeralopie. Klin Monatsbl Augenheilkd. 1928;81 :47 I - 473. 3. Mandcl IIG. Colin VH. Fat-soluble vitamins: vitamin A. In: AG Gilman. LS Goodman. TW Kail. F Murad, eds. The Pharmacological Basis of Therapeutics. 7th ed. New York: Macmillan; 1985: 1573-1582'. 4. Wilson JG, Roth CIJ, Warkany J. An analysis of the syndrome of malformations induced by maternal vitamin A deficiency: effects of restoration of vitamin A at various limes during gestation. Am J Anal. 1953:92:189-217. 5. Wei S, Episkopou V. Piantedosi R, et al. Studies on the metabolism of retinol and rciinol-binding protein in transihyretindcficicnt mice produced by homologous recombination. J ISiol Cheni. 1995;270:866-870.' 6. Dowling JE, Wald G. Vitamin A deficiency and night blindness. Proc Natl Acad Sci USA. 1958:14:648 - 66I. 7. Dhanda RP. Elcctroretinography in night blindness and other vitamin A deficiencies. Arch Ophthalmol 1955:54:841-849. 8. Sandberg MA, Rosen JU, Berson EL Cone and rod function in vitamin A deficiency with chronic alcoholism and in retinitis pigmentosa. Am J Ophthalmol. 1977:84:658-665. 9- Zrenner 1.:. Riither K. Apfelsiedt-Sylla I:. Retinilis Pigmentosa. Klinische Befundc, molekulargeneiische F.rgebnisse und Forschungsperspektiven. Ophthalmologe. 1992:89:5-21. 10. Fricdburg C, Sharpe I.T. licucl S, Zrenner E. A computer-controlled system for measuring dark adaptation and other psycho-

physical functions. Graefes Arch Clin ILxp Ophthalmol. 1998;

There has been little question that vitamin A deficiency is responsible for all xerophthalmic manifestations, from night blindness to corneal xerosis and for most, if not all, forms of xerophthalmic corneal ulceration.3''-35 However, the family presented here and the TTR-deficient mouse model demonstrate that the knowledge of the physiology of retinoid uptake and tissue delivery is far from complete. To summarize the syndrome, both affected siblings had no measurable serum RBP, one sixth of normal retinol levels, and normal retinyl esters. The RPE appeared morphologically and functionally to be the tissue affected most. Surprisingly, there were only mild or no changes in other organs, and there is evidence that most tissues can access retinyl esters present: in chylomicron remnants as an alternative to the retinolRBP-TTR complex. This may also be true for the lacrimal gland, because normal retinol levels were found in the tear fluid, which explains the lack of xerophthalmia in these patients. However, given the mandatory presence of RBP in the tear fluid and, during development, in the yolk sac, there is also evidence that there are different forms of RBP not affected by the genetic defect. It is a difficult task to find a treatment for this disorder. Although the changes to the retina and RPE exist probably for a long time, many investigators have shown that similar conditions caused by nutritive deficiencies respond fairlywell to vitamin A substitution. However, it is not trivial to deliver retinol to the RPE in the absence of its carrier. High doses of vitamin A are more likely to cause systemic side effects than to be useful. It appears as promising as challenging to generate and substitute synthetic or purified RBP.

236:31-40. 11. Marmor M, Zrenner II. Standard for clinical elecirorctinography (1994 update). Doc Ophthalmol. 1995:89:199 -210. 12. Marmor M. Zrenner I:. Standard for clinical clcctro-oculography. Arch Ophthalmol. 1993; 111:6() 1 - 604. 13- Sutler EE. Iran I). The field topography of ERG components in man, I. the phoropic luminance response. Vision Res. 1992;32: 433-446. 14. Seeliger M\V, Kretschmann Ull, Riither KW, Apfelstedt-Sylla E, Zrenner E. Muliifocal electroretinography in retinitis pigmentosa. Am J Ophthalmol. 1998; 125:214 -226. 15. Bicsalski UK, Frank J, Heck SC, el al. Rctinol-RHP is not needed to maintain all functions of vitamin A. Am J Clin Nutr. In press. 16. Seeliger MW, Kreischmann Ull. Apfelstedt-Sylla E, Zrenner E. Implicit time topography of muliifocal electroretinograms. bluest Ophthalmol Vis Sci. 1998:39:718-723. 17. Sommer A. Tjakrasudjatma S. Djunaedi E. Green WR. Vitamin A-responsive panocular xcrophihalmia in a healthy adult. Arch Ophthalmol. 1978;96:1630 -1634. 18. Bors I". Fells P. Reversal of the complications of self-induced vitamin A deficiency, lir J Ophthalmol. 1971:55:210-214. 19- Bavik C, Ward S|, Chambon P. Developmental abnormalities in cultured mouse embryos deprived of reiinoic acid by inhibition of yolk-sac retinol binding protein synthesis. Proc Natl Acacl Sci USA. 1996;93:3110-31 14. 20. Fishman GA. The electro-oculogram in retinal disorders. In: GA Fishman. S Sokol, eds. lUectrophysiologic Testing. San Francisco: American Academy of Ophthalmology; 1990:91-103. 21. Matsuo T, Matsuo N, Shiraga F, Koide N. Familial reiinol-bindingprotein deficiency. Lancet. 1987;2:4()2-4()3. 22. Episkopou V, Maeda S, Nishiguchi S, et al. Disruption of the transihyreiin gene results in mice with depressed levels of plasma retinol and thyroid hormone. Proc Natl Acad Sci USA. 1993;90: 2375-2379.

10VS, January 1999, Vol. 40, No. I 23- Wei S, Cpiskopou V. Piantedosi R. et al. Studies on the metabolism of rclinol and rciinol-binding protein in tr