The Pathology of Tangier Disease a Light and Electron Amicroscopic Study

Home , Abetalipoproteinemia, Hypobetalipoproteinemia, Sitosterolemia, Tangier disease

The Pathology of Tangier Disease A Light and Electron AMicroscopic Study

Victor J. Ferrans, MD, PhD and Donald S. Fredrickson, MD

Tangier disease (deficiency of high density plasma lipoprotems) is characterized clinically by: low levels of plasma cholesterol; enlarged, orange-yellow to yellow- gray tonsils and, frequently, penpheral neuropathy. Histologic and ultrastructural studies were made of various tissues from 5 patients with Tangier disease, and comparisons were made of these findings wih those in the 12 other patients thus far known to have this disease. Deposits of cholesteryl esters were found in: reticuloendothelial cells (foam cells) in tonsils, bone marrow, skin and jejimal submucosa; Schwann cells in peripheral nerves and myenteric plexus; and in nonvascular smooth muscle cells. These deposits appeared elecron lucent and intensely birefringent, varied from spherical to crystalline in shape, often were extensively confluent throughout large areas of cytoplasm, and were not limited by membranes. Certain foam cells in bone marrow also contained membrane- limited clusters of lipid particles resembling chylomicrons. The foam cells in Tangier disease differ morphologically from those in numerous lysosomal enzyme deficiency states, particularly Wolman's disease and cholesteryl ester storage disease, and in proliferative diseases of the reticuloendothelial system in which cholesteryl esters also accumulate in abnormal histiocytes. Morphologic and biochemical data suggest several hypothses to explain the accumulation of cholesteryl esters in tissues of patients with Tangier disease. Among these hypotheses, the most likely are considered to be the presence in plasma of abnormal lipoprotein particles that are subject to phagocytic removal by reticuloendothelial cells, and the failure of a process that normally removes locally synthesized cholesterol from cells to plasma. (Am J Pathol 78:101-158, 1975)

TANGIER DISEASE is a rare disorder of plasma lipid transport thought to be due to a mutant autosomal gene.' The homozygous abnormal patient has the following features: a) low plasma cholesterol and normal or elevated triglyceride concentrations; b) absence of the usual high density lipoproteins (HDL) and abnormal composition of other lipoproteins in plasma; c) enlargement and distinctive coloration of the tonsils; d) widespread tissue storage of cholesteryl esters, especially prominent in reticuloendothelial cells in lymph nodes, thymus, tonsils, bone marrow and rectal mucosa; and often, e) neuropathy. The combination of low plasma cholesterol and enlarged, yellow-

From the Section of Pathology and the Molecular Disease Branch, National Heart and Lung Institute, National Institutes of Health, Bethesda, Md. Accepted for publication September 10, 1974. Address reprint requests to Dr. Victor J. Ferrans, Section of Pathology, National Heart and Lung Institute, National Institutes of Health, Bethesda, MID 20014. 101 102 FERRANS AND FREDRICKSON American Journal of Pathology orange or yellow-gray tonsils or tonsillar remnants is considered pathognomonic of the disease.' Tangier disease was discovered in 1960 and given its name from Tangier Island, which lies off the coast of Virginia in the Chesapeake Bay and is the home of the original patients.2'3 As of this writing, 17 patients 2-14 have been discovered, representing 13 kindreds, none of which are known to be related. They include 10 Americans,2458 2 Germans,10 14 2 Australians,7'11'13 and 1 each from Switzerland,9 En- gland '7 and New Zealand.12 Fourteen have been described in previous reports,2-13 and 1 of these represented a presumptive diagnosis made at autopsy.13 Two patients are reported here for the first time (patients 15 and 16 in Tables 1 and 2). The seventeenth has been described to us in a personal communication,14 and is also included here. Some of the histopathologic features of Tangier disease have been described in reports of individual patients, but these have not been collated and reviewed. The ultrastructure of the foam cells has not been described in detail. This report provides light and electron microscopic observations in various tissues from 5 homozygotes with Tangier disease, as well as bone marrow studies in 1 obligate heterozygote, and reviews all other reported pathologic lesions in the context of current knowledge concerning the possible pathogenesis of Tangier disease. Case Reports Clinical and morphologic data on the 5 patients described in detail in this communication are summarized in Tables 1 and 2. All patients known to have Tangier disease are included in these tables, in which they are numbered according to chronologic order of diagnosis. This numbering system is also followed in the reports of individual patients and throughout the remainder of this communication.

Patient 2 ELa (CC No. 03-28-36), a white girl borm in 1954, was first seen at the Clinical Center of the National Institutes of Health in 1960. This patient and her brother TLa (patient 1 in Tables 1 and 2) were the first patients diagnosed as having Tangier disease. As reported in detail elsewhere;2'3 the detection of foam cells in the operatively removed tonsils of TLa led to a survey of his family on Tangier Island. ELa was asymptomatic and in good health except for enlarged, orange- yellow tonsils similar to those of her brother. The only abnormal findings disclosed by extensive laboratory studies were in plasma lipids and lipoproteins. Concentra- tions of plasma lipids, given in mg/100 ml of plasma and followed within parentheses by mean normal values for children,2 were as follows: total cholesterol, 85 (170), of which 23 (46) was in the form of free cholesterol; triglycerides, 237 (40), and total phospholipids, 82 (230). After paper electrophoresis of plasma, no stainable lipid was visible in the a-lipoprotein region. The high density lipoprotein fraction (density 1.063 to 1.21) contained only 2 mg/100 ml of cholesterol (mean Vol. 78, No. 1 TANGIER DISEASE 103 January 1975 normal for females, age 1 to 19 years, 56). Her tosils were removed, and they contained many large, foamy histiocytes similar to those in her brother's tonsils. Tonslar lipids, in mg/g dry weight of tissue, followed within parentheses by normal values 15 were: total lipids, 162 (69); cholesteryl esters, 89 (1); free cholesterol, 20 (1), and total phspholipids, 63 (49). All of the sterol present in the tonsils was cholesterol and the major (61%) acyl component of the cholesteryl esters was oleic acid.15 The patient has been followed regularly at the Clinical Center. In 1969 she developed left peroneal nerve palsy with foot drop. ITis disappeared, recurred for several months, and finaly subsided entirely. In 1972, she also develped transient right foot drop. Her plasma lipoprotein abnormalities have perssted. In July, 1971, her plasma cholesterol was 62 mg/100 ml, and the triglycerides were 186 mg/100 ml; cholesterol concentrations in lipoprotein fractions were (in mg/100 ml of plasma, followed by normal values, ie, 90% fiducial limits for females ages 1 to 19 years,16 given within parentheses): very low density lipoproteins (VLDL) 17 (5- 25); low density hpoproteins (LDL) 44 (50-170); and high density lipoproteins (HDL), 1 (40-5). Examination of bone marrow (1961) by phase contrast and polarization microscopy revealed foam cells resembling those in her tonsils. A few foam cells with similar characteristics were found in bone marrow of her mother, but not of her father. Histologic study of punch biopsy (1965) of her slin failed to demon- strate the lipid-contaiing cells which have been found4 in the dermis of other patients with Tangier disease. Ught and electron microscopic observations on foam cells observed in the lamina propria of jejunal mucosa obtained by peroral biopsy in 1972 are described in this report. Patient 4 PeLo, (CC No. 05-32-90-3), a 19-year-old white girl, was diagnosed as having Tangier disease at age 7, when she had a tonsillectomy for greatly enlarged, orange tonsils. The tonsils were shown to contain deposits of cholesteryl esters and the patient was referred for evaluation to the Clinical Center of the National Institutes of Health. The patient was found to have a low plasma cholesterol and absence of a-lipoproteins. She also had congenital pulmonic valvular stenosis, a lesion which was operatively corrected when she was 10 years old. At that time, extensive deposits of cholesteryl esters were found in foam ceHls in the thymus and lymph nodes; small, focal deposits also were found in the wall of the pulmonary artery; atrial myocardium appeared free of lipid deposits.15 The patient did well postoperatively and has had no significant cardiac symptoms since then. She developed insulin-dependent diabetes mellitus at age 14 and has had adequate control of this disease. She has no signs or symptoms of peripheral neuropathy and no hepatosplenomegaly. Her sister (patient 3 in Tables 1 and 2) also has Tangier disease. When last seen at the Clinical Center, in November of 1973, her plasma cholesterol was 101 mg/100 ml; triglycerides were 178 mg/100 mL Cholesterol concentrations in plasma lipoprotein fractions (in mg/100 ml of plasma) were: VLDL, 19; LDL, 73; and HDL, 9. A biopsy of chnically normal sin was taken for ulrastructural study. Patint 5 CNo, (CC No. 04-93-64-8), a 55-vear-old man, was first diagnosed as having Tangier disease in 1962,5 when he underwent evaluation of splenomegaly, mild anemia and intermittent diarrhea of several years' duration. His tonsils had been removed at age 14 because of marked enlargement. Physical examination disclosed fine comeal deposits such as those seen in other adults with Tangier disease,5 104 FERRANS AND FREDRICKSON American Journal of Pathology

U) = ~~ 0CL 4--0a )-( 3: :p C) (n E uko ow. E E.~~E~L~~ ~ 0 a-4*- bt 0 a) 4.O '0 (U C )(-( E 0 E -c

M CL4-1- 0 W) E3 :6

+ +o o + + ++ + o ++ 0 + 0 o 76) a- 0 0 -a)0 I + + + I + + I ++ o a1) ' C 0 nU) 0tn (Ua) Eo a) U) + + + +0 + + . _ +0 o ' Q w to (U) OE o a) -c O + + D C :t + + I + m0 - C s 0amE U) 2O O + 0 1 1 CU - a) + 000 + 03 a) (L E r-.

'-I ._ C C - + ++ +0 + + + + + + + U) on ._ . ) M x > > C C mU o iz + + 0+ + I I - C m O cn c0 U) '4 C) ' o >ZN a) I' .0 Le) CD " r. C00 I_ CD -fr-. S 04. _ '-4 lq lq m r- c" qr lq ul m Le) . U) cn , U) ca C 0 0 Q .' X)a)U) Z .0 0 ._ 2 LL. Lu L L UL. U- 4. a)-0 (U

- In) Lfl 0 C 9-4..- (U U) 4. a) r-. 0~~~~~~~~~~~- s -- U) o c%JC".Z fl) Lfl tIn Coo co r P. (a a) ) HL.. H2 (U

I " m q 10 U, r-_ oOO 9.-4 9-4 9-4 9-4 9-4 v-4 Vol. 78, No. 1 TANGIER DISEASE 105 January 1975

0 E '.n 0 .-0 0C > Eo -Co _ _ C 0 0 _ ; .- aU- o e 0.> ~. - s - E a a > E% 0 - 0 o 000 o CL~ 0to 0 o X c c -3c ' E' I E °E Uo -, i0 C C 0 z cn (aQ -_) 0a E S- I + I ++ I +10 0 I I ++ II + I +O+ tna + + I I I° 0 + I I + I I + I

I I I I +++ I + I I++ S coSo I I I I E C .i.0E 0 C S I + I 100 I I I + SC S 4- I II I I I I to o a c S on C 0 Cb. Sc aa ++O I I I I +l l+ I I 0 0 I I I U) cL z aW- E 0 0 0 C I++II I 0 + I 10 cL I++ I

1-r... .0 C I I I I I I I+ I I I I a_ C E a C C n I - S i-) + II ++ I I I I I I I+ .0I _ 0 I I I I U 0.a C SC 0 0 +1+1+ I 1+1+ 4n c ,0I -C - ilf C" "o so cmlI SLI K ,-i aOeo- 0 (%JC ° . osr er i in ~a c,0o cm°W-i rN V I-

- cmJ C) Mr- 4o r-.. co OD 0CO (%J Z -4 -4- -- D- 9- 106 FERRANS AND FREDRICKSON American Journal of Pathology patches of yellow-orange lymphoid tissue in the pharynx, and hepatosplenomegaly. He had no lymphadenopathy, cutaneous lesions or neurologic abnormalities. Laboratory studies disclosed leukopenia, thrombocytopenia, mild anemia and reticulocytosis. Plasma cholesterol was 38 mg/100 ml; triglycerides were 142 mg/100 ml. The cholesterol content of plasma lipoprotein fractions (in mg/100 ml of plasma) was: VLDL, 19; LDL, 11; and HDL, 1. Slit lamp examination showed random, soft densities involving the entire stromal thickness of each cornea. Roentgenograms of the colon, small intestine and stomach were negative, but proctoscopy disclosed the presence of multiple, tiniy, flat, orange-brown spots throughout the rectum. Biopsy of these areas showed large, foamy histiocytes in the mucosa and submucosa. Foci of foam cells also were found on biopsy of liver and bone marrow. Biopsy of jenunal mucosa revealed no abnormalities. The patient's gastrointestinal symptoms responded to a low-fat diet. In 1963, the patient underwent splenectomy for relief of continuing hypersplenism and discomfort caused by increasing splenomegaly. The spleen was firm, greatly enlarged (1,160 g) and contained numerous foam cells; the liver appeared only slightly enlarged and was grossly normal. During the postoperative period the patient developed an acneiform eruption over the trunk and abdomen. This eruption consisted of 2- to 3-mm papules, many of which were perifollicular. Biopsy of these lesions revealed follicular plugging and granulomatous foci containing lymphocytes, multinucleated giant cells, fibroblasts and macrophages laden with birefringent lipid deposits. Since 1963, these papules have continued to appear and disappear spontaneously, but in much smaller numbers than at the time of their onset. When the patient was seen at the Clinical Center 8 months after operation, it was noted that his plasma cholesterol concentration had increased to 155 mg/100 ml without change in his extremely low level of HDL. Before operation, numerous determinations had shown his plasma cholesterol concentration to be consistently below 50 mg/100 ml. The patient was admitted again to the Clinical Center in 1966, at which time he was found to have wasting and weakness of the gastrocnemius and infraspinatus muscles bilaterally. Nerve conduction times were decreased; fasciculations were observed on electromyography, and a calf muscle biopsy showed evidence of denervation. In 1972 he experienced the onset of a sensory neuropathy involving the left median nerve. When last seen at the Clinical Center, in October of 1973, he showed no change in the corneal opacities or in the degree of hepatomegaly, but he had developed glucose intolerance. Little progression of his peripheral neuropathy was evident. His plasma cholesterol was 101 mg/100 ml and the triglycerides were 362 mg/100 ml. The choesterol content of plasma lipoprotein fractions (in mg/100 of plasma) was: VLDL, 36; LDL, 65; and HDL, 0. A biopsy of skin from the upper abdomen was taken for ultrastructural study.

Patient 15 JSt, a 6-year-old boy (CC No. 09-35-67-0), was first seen at the Clinical Center of the National Institutes of Health in July, 1972. In April, 1972, his enlarged tonsils were removed because of recurrent ear infections, and the diagnosis of Tangier disease was suspected from histologic studies. The boy had been adopted and no medical information conceming his blood relatives was available. At 11 days of age he required operation for relief of hypertrophic pyloric stenosis, and since age 2 he had undergone a series of corrective operations for recurrent im- balance of external ocular muscles. His growth and motor development were considered normal, but he was thought to have borderline mental retardation. On physical examination he was found to have mottling of retinal pigment epithelium Vol. 78, No. 1 TANGIER DISEASE 107 January 1975 in the periphery of the retina and around the macula in both eyes, changes not observed in other patients with Tangier disease. The corneas were clear. The possibility of oculmmotor nerve palsies could not be fully evaluated because of the effects of the previous operative interventions. The spleen, liver and lymph nodes were not enlarged. No abnormalties of the peripheral nervous system were found. Results of audiogams and studies of color vision, dark adaptat and reftinal cone thresholds were normaL Radiographic survey of chest, abdon and bones was considered normal for age. Hemogram, urinalysis and routi clinical chemical analyses were within normal limits. The plasma cholesterol level was 51 mg/100 ml; that of triglycerides was 97 mg/100 mL Concentrations of cholesterol in plasma lipoprotein fractions (in mg/100 ml Of psma) were: VLDL, 1; LDL, 48; and HDL, 2. Bone marrow aspiration revealed foam cells similar to those found in other patients with Tangier disease. PaienLt 16 RJa (CC No. 09-70-10-4), a white boy bom in 1970, underwent tonsillectomy and adenoidectomy in September, 1972, for relief of chronic mouth breathing, nasal obstruction and recurrent otitis media. Te diagnosis of Tangier disease was suspected on the basis of the typcal gross and histologic appearance of his tonsils, and was subsequently confirmed by studies of plasma lpoproteins. The patient was first seen at the Clinical Center of the National Institutes of Health in Aprit 1973, at which time he had no symptoms attributable to Tangier disease. Physical examination disclosed recurrent adenoidal tissue, and a smooth, firm and non- tender liver that was palpable 2.5 cm below the right costal margin; the tip of the spleen also was palpable. Detailed neurologic and opthamologic inations revealed no abnormalhties. Total plasma cholestrol was 83 mg/100 ml and plasma triglycerides, 207. Cholesterol content of plasma lpoproteins (in mg/100 ml of plasma) was: VLDL, 33; LDL, 50, and HDL, 0. Results of other extensive laboratory studies were within normal limits. A bone marrow aspirate, taken for light and eletron microscopy, revealed large numbers of foam cels. A biopsy of clinically normal slin was taken for utrastructural study. The patienes father, PJa (CC No. 09-70-11-6), also was evaluated at the Clinical Center in 1973. He was 33 years old, had a history of allergic asthma, and underwent tonillectomy for recurrent tonillitis at age 20; he had no signs or symptoms attributable to Tangier disease. Physical examination disclosed no abnormalities. His plasma cholesterol was 142 mg/100 ml; plasma triglycerides were 115 mg/100 ml. Cholesterol content of plasma lipoptein actions (in mg/100 ml of plasma) was: VLDL, 23; LDL, 91; and HDL, 28. Examination of bone marrow revealed small number of foam cells similar to those in other patients with Tangier disease. He is considered to be an obhigate heteroygote for Tangier disease.

Materials and Methods L;ght Microscopic Stdies Parts of the jejunal biopsy specimen from patient 2, the tonsils of patients 2 and 15, and skin from patients 4, 5 and 16 were fixed in 10X formalin and embedded in paraffin for preparation of histoogic sections, which were stained with hema- toxyhn and eosin. Frozen sections of other portions of these tissues were stained with oil red 0, Baker's, Sudan black B, periodic acid-Schff, alcian blue (pH 2.5) and Rinehart stains. Aspirates of bone marrow were obtained from patients 15 and 16, and from the father of patient 16. Portions of each sample were imediately 108 FERRANS AND FREDRICKSON American Journal of Pathology examined, without any fixation or staining, by phase contrast, Nomarski interference contrast and polarization microscopy. Other portions were used for the preparation of smears, which were stained according to Wright's method, and for ultrastructural study. Electron Microscopic Studies For electron microscopy, part of the jejunal biopsy specimen from patient 2, skin biopsies from patients 4, 5 and 16, and fragments of bone marrow from patients 15 and 16 and from the father of patient 16, were fixed with cold 3% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2. After washing with several changes of 5% sucrose in 0.1 M phosphate buffer, pH 7.2, the tissues were fixed with cold 1% osmium tetroxide in Millonig's phosphate buffer, dehydrated and embedded in Maraglas.17 Tonsillar tissue from patient 15 had been previously fixed with phosphate-buffered formalin. This tissue was refixed with glutaraldehyde and further processed as described above. Semifine (0.5 u thick) sections from all tissues were stained with alkaline toluidine blue and examined by light microscopy. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with the electron microscope.

Results Light Microscopic Observations Although they varied in size from one tissue to another, the foamy histiocytes appeared morphologically similar in jejunum of patient 2 (Figures 1-5), tonsils of patients 2 and 15 (Figures 6-10), bone marrow of patients 15 and 16 (Figures 10-15) and in skin of patients 4, 5 and 16 (Figures 16-18). As described in detail below, foamy vacuolization of the cytoplasm was found also in Schwann cells and smooth muscle cells.

Jejunal Biopsy The foamy histiocytes in the jejunal biopsy from patient 2 were located in the submucosa and were relatively small (15 to 30 p in diameter). These cells (Figures 1-5) were rounded or oval in shape and had one or two small, darkly staining nuclei which often were indented by lipid in the cytoplasm. Numerous vacuoles, ranging from 1 to 5 [t in diameter, were present throughout the cytoplasm. This vacuolization was better appreciated in semifine sections of plastic- embedded tissues (Figure 5) than in paraffin sections stained with hematoxylin and eosin (Figure 2). In frozen sections stained with oil red O or with Sudan black B it was evident that the cytoplasmic vacuoles seen in extracted preparations were lipid deposits (Figure 3). Most of this lipid exhibited intense birefringence of the crystalline type (Figure 4). Only a few deposits showed the Maltese cross type of birefringence. The lipid deposits did not stain with the periodic acid- Vol. 78, No. 1 TANGIER DISEASE 109 January 1975 Schilf, Baker, Rinehart or alcian blue methods. A few, small granules which showed yellow-green autofluorescence and stained with the periodic acid-Schiff method were present in many foamy histiocytes. These granules were considered to be ceroid. Smooth muscle cells in the muscularis mucosa (Figure 2) and Schwann cells (Figure 5) showed variable degrees of vacuolization in lipid-extracted preparations. The architecture of the jejunal villi appeared normal. No foam cells were present with villi, although small numbers of extracellular lipid droplets were found throughout the mucosa. No inflammatory reaction was associated with the foam cells.

Tonsils The foam cells were largest in the tonsils (Figures 6-10), where they reached diameters of 70 [ (Figure 10). The foam cells in the tonsils of both patients (2 and 15) were located primarily in the fibrous septa (Figures 6 and 7) but also were present in small numbers within the follicles (Figure 6). The histochemical characteristics of foam cells in tonsils and jejunum were similar. Examination by polarization microscopy of sections stained with oil red 0 revealed that some of the tonsillar lipid deposits did not show any birefringence (Figures 8 and 9); furthermore, some large crystals stained poorly with oil red 0 or with Sudan black B. Birefringent crvstals were generally larger in tonsils than in jejunum (Figures 8 and 9).

Bone Marrow Foam cells in bone marrow of patients 15 and 16 and the father of patient 16 were easily identified in semithin sections of plastic- embedded tissues (Figure 11) and in fresh, unfixed spreads examined by bright field (Figure 12), polarization (Figure 13), phase contrast (Figure 14) and Nomarski interference contrast (Figure 15) microscopy. The cytoplasm of these cells contained refractile lipid droplets which exhibited the Maltese cross type of birefringence (Figures 13 and 14). The shapes of the foam cells varied considerably; some cells were round, while others were slender and elongated. Most foam cells were smaller (15 to 30 t in diameter) than those which we have observed in type 1 hyperlipoproteinemia 8'19 and in Niemann-Pick disease.20 In fresh preparations they also seemed much more fragile and often appeared to have ruptured, releasing their birefringent droplets into the extracellular environment. The foam cells in bone marrow of the father of patient 16 were 110 FERRANS AND FREDRICKSON American Journal of Pathology smaller, appreciably less numerous and much more intensely autofluorescent than they were in patients 15 and 16. They were similar, however, in their content of birefringent material and in their pleomorphic shapes.

Skin The abnormalities observed in biopsies of clinically normal skin from patients 4, 5 and 16 consisted of extensive vacuolization of the cytoplasm of Schwann cells in small, unmyelinated, cutaneous nerves, and of histiocytes. The changes in Schwann cells were clearly demonstrable in semifine sections of plastic-embedded tissues, but were not discernible in ordinary histologic preparations. Cutaneous blood vessels appeared free of lipid deposits.

Electron Microscopic Observations Jejunum Histiocytes, smooth muscle cells and Schwann cells were the obvious sites of lipid accumulation in the jejunum. The size and ultrastructural features of these types of cells varied according to the degree of lipid deposition which they exhibited. The foamy histiocytes were located subjacent to the smooth muscle cells of the muscularis mucosa and were surrounded by collagen fibrils, microfibrils and finely granular to amorphous ground substance (Figures 19-21). Their surfaces showed moderate to marked irregu- larities of contour, usually in the form of invaginations or of pocket- like flaps which extended into the interstitium. Their basement membranes were either absent or very poorly developed. These cells were uni- or binucleated. Their nuclei were indented by lipid deposits in the cytoplasm (Figure 19). The nuclear chromatin was finely dispersed throughout, with some tendency to margination. Nuclear pores and nucleoli were small and inconspicuous. The following structures were identified in the cytoplasm of the foamy histiocytes: mitochondria, Golgi zones, glycogen particles, free ribosomes, centrioles, rough- and smooth-surfaced endoplasmic reticulum, microtubules, 100-A-thick filaments, pinocytotic vesicles, residual bodies, lysosomes and lipid droplets. The relative proportions of these components varied considerably in different cells (compare Figures 19, 20 and 21). In cells which were heavily laden with lipid deposits, other cytoplasmic structLres were markedly reduced in number and were distributed in narrow bands or small polygonal zones between Vol. 78, No. 1 TANGIER DISEASE 111 January 1975 closely adjacent lipid deposits (Figures 19 and 21). The great majority of the foam cells present in the biopsy specimen were of this type. Other cells were less severely involved and contained more abundant cytoplasmic organelles; their perinuclear zones tended to be free of lipid deposits. Golgi areas, free ribosomes, 100-A filaments and elements of smooth- and rough-surfaced endoplasmic reticulum were typically present in these perinuclear regions. The mitochondria in foamy histiocytes were few, small, and contained sparse numbers of cristae and electron-dense intramitochondrial granules. The 100-A-thick filaments followed straight or slightly wavy courses between cytoplasmic organelles (Figures 22 and 23) and frequently formed meshworks around the lipid deposits (Figure 23). The Golgi zones were prominent in cells with few lipid deposits (Figure 20) but small and inconspicuous in other cells. These zones consisted of several stacks of closely apposed, flattened, elongated cisterns which were associated with numerous vesicles, smooth-walled tubules, small lysosomes, and pleomorphic residual bodies of various sizes. All of these structures were limited by single, trilaiinar membranes. The vesicles measured from 500 to 1,000 A in diameter and their content varied from lucent to markedly dense. The small lysosomes and many of the smooth-walled tubules were filled with amorphous, electron- dense material (Figure 22). The residual bodies (Figure 22) contained electron-dense material which varied from amorphous to lamellar in structure. Large, sometimes markedly elongated crystal clefts were found within many of the residual bodies (Figure 22). Some of the lamellae in residual bodies showed a concentric arrangement with irregular periodicity. Other residual bodies contained lipid droplets surrounded by electron-dense material. The electron-dense material within residual bodies was judged to correspond to that which was autofluorescent and periodic acid-Schiff-positive in histologic preparations. Its ultrastructure was similar to that of ceroid deposits which we have described in foam cells in type 1 hyperlipoproteinemia.18'19 Because of these features, this material was considered to represent ceroid. Most of the lipid deposits in foamy histiocytes were not limited by membranes and appeared free in the cytoplasm. Some of these deposits were composed of faintly stained peripheral zones and lucent centers (Figure 21); however, the great majority of deposits were completely electron lucent. This indicated that their content had been extracted during preparation of the tissues for electron microscopy. In many cells the lipid deposits appeared to have coalesced, 112 FERRANS AND FREDRICKSON American Journal of Pathology forming large, multilobulated, electron-lucent areas. Many of the lipid deposits in the most severely involved cells (Figure 21) had coalesced to such an extent that they appeared continuous throughout large areas of cytoplasm. Most of these large, coalescent lipid deposits were completely electron-lucent and irregular or polygonal in shape. In contrast to this, lipid deposits in less severely involved cells were rounded and noncoalescent. Many of the lipid deposits were demar- cated from adjacent areas of cytoplasm by thin (25 to 50 A in thickness), single or multiple electron-dense layers which did not have the trilaminar structure of true membranes (Figures 23-26). Other lipid deposits were not separated from the rest of the cytoplasm by any intervening structures. No morphologic evidence of lipid uptake was observed in the foamy histiocytes, and no lipid particles were observed in their vicinity. Some lipid deposits which by light microscopy appeared to be in extracellular spaces were actually located within elongated cytoplasmic processes of foamy histiocytes. Foamy histiocytes were not found in the mucosal surfaces or in the villous cores, which appeared ultrastructurally normal. No lipid deposition was found within vascular smooth muscle or endothelial cells, although electron-lucent areas which could have represented lipid deposits were present in extracellular areas of vascular walls. Foamy histiocytes often formed one or more layers external to the smooth muscle cells of the walls of blood vessels (Figure 27). In contrast to vascular smooth muscle, smooth muscle cells in the muscularis mucosa often contained lipid accumulations (Figure 28). The amount of lipid varied from one cell to another. Some cells appeared distended and had large, central cores composed of lipid droplets and poorly organized lamellar material; these structures had displaced the contractile elements toward the cell periphery, where they were arranged in a narrow band subjacent to the plasma membranes (Figure 28). By light microscopy these cells were difficult to distinguish from foamy histiocytes. A consistent finding in areas of involvement of smooth muscle cells was the presence of electron- lucent lipid particles in extracellular spaces (Figure 28). These particles measured from 1,000 to 3,000 A in diameter and were enmeshed in the basement membranes and connective tissue fibrils between adjacent smooth muscle cells, often in close proximity to their plasma membranes. No evidence was observed of actual engulfing of lipid particles by smooth muscle cells. Numerous nerves and associated Schwann cells were identified in Vol. 78, No. 1 TANGIER DISEASE 113 January 1975 the submucosal layers. The Schwann cells consistently were the sites of accumulation of numerous lipid droplets of low electron density (Figures 29 and 30). These droplets also were present in areas of Schwann cell cytoplasm surrounding unmyelinated axons. Many axons contained small, electron-lucent areas which probably represented lipid droplets.

Tonsils The foam cells in the tonsils of patient 15 showed the same ultrastructural characteristics exhibited by the foamy histiocytes in the jejunal submusosa of patient 2. The foam cells in tonsils (Figure 31) were larger, however, and contained mostly the extensively coalescent type of lipid deposits.

Bone Marrow The foamy histiocytes in bone marrow of patients 15, 16 and the father of patient 16 showed the greatest degree of pleomorphism in foam cells in Tangier disease. As in other organs, the basic feature of foam histiocytes in bone marrow was the presence of a) numerous, large, electron-lucent lipid deposits which were not limited by trilaminar membranes, and b) residual bodies of variable electron density. Cells which contained primarily these two types of structures were considered to be fully developed, lipid-laden histiocytes (Figure 32). In addition to moderate amounts of lipid, a few histiocytes contained phagocytized erythrocytes in various stages of digestion (Figure 33). Other histiocytes (Figures 34-6) differed structurally from those described above and were considered of special interest in that they offered a basis for speculating on the mechanisms of formation of foam cells. These histiocytes were characterized by their content of a) large (2 to 31L in diameter) pale gray lipid droplets which were free in the cytoplasm, and b) clusters or aggregates of smaller (1,000 to 5,000 A in diameter) lipid particles which often appeared partially coalescent (Figures 34-36). Some of the latter aggregates were free in the cytoplasm (Figures 34 and 35) while others were enclosed within vacuoles limited by single, tilaminar membranes (Figure 36). The particles within these vacuoles were embedded in electron-dense matrices composed of amorphous material and small, irregularly spaced lamellae (Figure 36). Numerous, minute, electron-lucent areas were present within some of the lipid particles which were free in the cytoplasm (Figure 35). The foamy histiocytes in the bone marrow of the father of patient 114 FERRANS AND FREDRICKSON American Journal of Pathology 16 exhibited the same range of morphologic features as described above, including the aggregates of small lipid particles (Figures 37 and 38). These cells differed ultrastructurally from those in bone marrow of patients 15 and 16 in their content of numerous ferritin- like particles and of pleomorphic residual bodies.

Skin Ultrastructural study confirmed the light microscopic observation that Schwann cells in small, unmyelinated nerves (Figures 39 and 40) were the sites of extensive lipid deposition in the skin of patients 4, 5 and 16. These deposits appeared similar to those observed in neural elements in jejunum of patient 2. They involved practically all Schwann cells encountered and were associated with the presence of unusually large numbers of collagenous fibrils separating adjacent axons and Schwann cells in small nerves. A few axons contained focal accumulations of glycogen particles and/or concentric lamellae (myelin figures); others exhibited small, lucent areas, similar to those in axons in jejunum of patient 2. In the skin of patients 4 and 5, the- Schwann cells and the capillary endothelial cells had thickened basement membranes. These changes were more pronounced in patient 4 (who has insulin-dependent diabetes mellitus) than in patient 5 (who has glucose intolerance but not clinically overt diabetes mellitus). Involve- ment of Schwann cells was least pronounced, but still clearly discernible in patient 16, the youngest patient known to have Tangier disease. Capillary endothelial cells and smooth muscle cells in cutaneous blood vessels were free of lipid deposits, but histiocytes in perivascular locations showed lipid accumulations resembling those in histiocytes of other organs.

Discussion The clinical and pathologic features of the 17 known patients with Tangier disease are summarized in Tables 1 and 2. The typical changes in plasma lipids and lipoproteins were demonstrated in 16 of these patients (1-12 and 14-17), while the diagnosis in patient 13 was made only at autopsy. The latter patient 12 apparently incurred cerebral trauma at birth, had severe spasticity and mental retardation, and died of bronchopneumonia at 5% years of age. His parents had concentrations of plasma HDL cholesterol (20 and 13 mg/100 ml) low enough to be compatible with the presumed heterozygous phenotype for this disease.' One other patient with Tangier disease (6) is known Vol. 78, No. 1 TANGIER DISEASE 115 January 1975 to have died. This patient had symptomatic coronary artery disease and died suddenly.5 Autopsy was not performed. Before we proceed to generalizations concerning the pathologic findings in Tangier disease, to compare them with findings in certain other lipid storage diseases, and to engage in speculations about pathogenesis, it may be helpful to review the known biochemical features of the disease.

Biochmical Feares of Tangie Dse Although Tangier disease is clearly inheritable, the genetic locus or basic biochemical defect is not established, and it cannot be assumed that all patients with tis syndrome necessarily represent an identical mutation. Extensive chemical analyses have established that the tissues laden with foamy histiocytes contain excessive amounts of cholesterol (5-cholestene-3i-ol) in the form of esters. The most common of the acyl moieties is oleic acid, and the general fatty acid composition closely resembles that in the cholesteryl esters normally found in plasma or deposited in other diseases in which tissue concentration of such esters is increased (see Frederickson et al,1 Frederickson 5 and Sloan and Frederickson21 for reviews). There is no accumulation of oxygenated sterols, such as has been recently found in Wolman's disease,22 or of cholestanol or plant sterols, as may occur in other genetic disorders of sterol metabolism.2'24 Furthsermore, there is no increase in concentration of any of the tissue phospholipids or sphingo- lipids. A modest increase in triglycerides has been present in a few isolated tissues from patients with Tangier disease,"',5 but this is not a consistent abnormality. Most work on the biochemical defect has centered around the peculiar changes in plasma lipoproteins. Ihe published data concerning the lipoproteins in Tangier disease 1"25 have been supplemented by quite recent studies of large quantities of plasma from a number of such patients.'28Most attention has been devoted to the HDL. In the analytical ultracentrifuge, no lipoproteins of the HDL density class are seen.25 Only very small quantities of cholesterol, one-tenth to one-fiftieth of the normal amount, -are captured in the preparative ultracentrifuge in the usual HDL range. The small amounts of lipoprotein in this fraction have been referred to as "HDLr" (Tangier HDL). It has been previously reported that HDLr was particularly deficient in one of the major apolipoproteins,29 but the more recent data have considerably extended and further clarified the nature of the abnormality.2628 116 FERRANS AND FREDRICKSON American Journal of Pathology

High Density Lipoproteins Currently, there are considered to be four major families of plasma lipoproteins.-" They contain practically all of the circulating lipids in plasma with the exceptions of: the free fatty acids, which are bound to albumin; vitamin A and possibly others of the fat-soluble vitamins, for which specific transport proteins exist; and small amounts of phospholipids, mainly lysolecithin, which are bound to proteins in complexes having a density greater than 1.21 g/ml. The lipoproteins have a density of less than 1.21, which permits their isolation from other plasma proteins by ultracentrifugation in solutions raised to this density by the addition of salt. Three of the four lipoprotein families, namely, chylomicrons, VLDL and LDL, appear to be mainly concerned with triglyceride transport. LDL arise mainly from degradation of chylomicrons and VLDL by a process which mainly involves removal of triglycerides from these particles as they pass along the endothelial cells. Two enzymes are believed to have important roles in this process. One is lipoprotein lipase, which catalyzes hydrolysis of the ester bonds in triglycerides. The other is lecithin:cholesterol acyltransferase (LCAT), which catalyzes transfer of a fatty acid from the 2 or 1 position of lecithin to cholesterol, thus forming cholesteryl esters. It has been postulated, but not proved, that the LCAT reaction is important during hydrolysis of chylomicrons and VLDL, replacing departing glycerides with suffi- cient cholesteryl esters to stabilize these particles as they diminish in size.31 HDL, or a-lipoproteins, are not unconcerned with these catabolic processes. They contain apolipoproteins that activate lipoprotein lipase (apoC-II) and LCAT (apoA-I). They also provide a source of lecithin for LCAT-catalyzed esterification and may provide a "sink" for some of the product cholesteryl esters. It is generally believed, however, that HDL have other, yet undiscovered physiologic roles. Teleologically, HDL may be the most primitive lipoproteins, for in many species they carry the bulk of the lipids in plasma. In the human fetus, they also are the major lipoprotein class, but after birth the HDL concentration is quickly exceeded by the combined concentrations of VLDL plus LDL. HDL have a density lying between 1.063 and 1.21 g/ml. They are spherical particles having a diameter of between 80 and 150 A, and in overall composition are about half protein and half lipid by weight. About 20% of the dehydrated mass of HDL is phospholipid, 20% cholesteryl esters, and the remaining lipid consists of small amounts Vol. 78, No. 1 TANGIER DISEASE 117 January 1975 of triglycerides and unesterified cholesterol. The protein complement of the whole HDL fraction in man consists of about 70% apoA-I, 20% apoA-II and 5% low molecular weight C-apoproteins. The A apoproteins in plasma are almost entirely confined to HDL,," although small amounts occur in chylomicrons.3- The mass of C-apoproteins is about equally divided between VLDL and HDL. Evidence is accumulating that at least a significant portion of the C-proteins are secreted in HDL and are made available to VLDL as the latter particles are secreted from liver and small intestine. The amino acid sequence of apoA-II in HDL is sufficiently close to the primary structures of both apoC-I and apoC-III to suggest that all three of these proteins probably arose from a common ancestral gene.-3 These and other data raise many questions, which have not yet been answered, about the rela- tionships of HDL to secretion and metabolism of the triglyceride-rich classes of lipoproteins. The structure of normal HDL also has not been fully elucidated. The present information supports apoA-1I as being a more avid binder of lipid than the larger and more abundant apoA-I molecule.36 Sug- gestions have been made that HDL contains a "core" of apoA-II plus lipid, to which apoA-I and the C-proteins are attached mainly by protein-protein interactions.'

Tangier HDL The importance of these latter suggestions has perhaps been increased by the recent finding of tiny amounts of this possible A-II "core particle" in Tangier plasma.Y6 Small amounts of apoA-I and apoA-II are present in plasma of all patients with Tangier disease. All of the apoA-I is present in the plasma fraction of density greater than 1.21 g/ml.Y In the HDL density range, small, spherical lipoprotein particles, about 60 A in diameter, have been repeatedly isolated from plasma of three Tangier homozygotes. These lipoproteins (A-II particles) contain only apoA-II along with a lipid complement indistinguish- able from that of whole normal HDL. At this time, the genetic defect in Tangier disease seems most likely to be some alteration in apoA-I, perhaps a structural anomaly, that prevents integration of this apoprotein into the usual quaternary structure of HDL. The result is an inability to maintain an effective concentration of HDL, including loss of the usual C-protein reservoir provided by the HDL family of lipoproteins. There are also peculiarities in the composition and, therefore, possibly in the structure of the other plasma lipoproteins in Tangier disease. 118 FERRANS AND FREDRICKSON American Journal of Pathology Firstly, there is an apparent reduction in cholesterol-carrying capacity in plasma beyond that due to simple loss of HDL. LDL concentrations are quite low, sometimes a tenth of normal, and the composition of LDL, particularly of the usually more concentrated S, 0-8 class (density 1.019 to 1.063), is very low in cholesterol and excessively high in triglycerides.27 The B-apoprotein or normal protein constituent of LDL is apparently in adequate supply, however, and the LDL levels are comparable to those seen in familial hypobetalipoproteinemia.1 Neither this latter syndrome nor Tangier disease are accompanied by the malabsorption, acanthocytosis, retinitis pigmentosa or posterior col- umn lesions characteristic of abetalipoproteinemia (total deficiency of B-apoprotein) .' Secondly, although capability for transport of triglycerides in plasma in chylomicrons and VLDL is not abolished in Tangier disease, as it is in abetalipoproteinemia, triglyceride metabolism is not entirely normal. The several C-apoproteins, which together with apoprotein B are the principal protein components of chylomicrons and VLDL, are all elaborated. Normally, about half of the C-apoproteins reside with the HDL, however, and appear to shuttle between HDL and the triglyceride-rich particles to stabilize the latter and to promote their metabolism.30 Apoprotein C-II, for example, activates the lipolysis catalyzed by lipoprotein lipase during removal of chylomicrons and VLDL. In Tangier disease, total plasma triglycerides usually are too high. Chylomicrons appear to leave plasma rather slowly and their content of cholesterol is abnormally low."5 Compositional analyses of Tan- gier VLDL have revealed a relative deficiency of C-protein in these particles. This deficiency may be related to the peculiar migration of Tangier VLDL on paper electrophoresis.'5 When the unusual "A-II particles" that constitute the HDL in plasma of Tangier homozygotes are isolated in the ultracentrifuge, they are accompanied by large, disc-like lipoproteins of a diameter of about 1,000 A which can be separated from the A-II particles by chromatog- raphy.26 The very large particles are present in greater concentrations when the patients are eating fat, and may be absent when all fat is removed from the diet.27 In accordance with an older theory,2 it is possible that these large, peculiar particles are remnants of chylomicrons which, because of abnormalities either in their original composition or in their subsequent metabolism, are unstable and become "foreign bodies" subject to phagocytosis by histiocytes. Estimations of plasma LCAT activity by in vitro measurements have led to conflicting conclusions in 2 patients with Tangier disease.13'14 Activity was within normal range in 1 patient 14 and reduced in the Vol. 78, No. 1 TANGIER DISEASE 119 January 1975 other.13 The amount of lipoprotein lipase activity transiently displaced into plasma by heparin has been considered normal in 1 patient 14 and low in another.13 Several other homozygotes have been shown to have abnormally low responses to this test in studies from this laboratory.37 It has been possible to measure cholesterol synthesis only once in Tan- gier disease, specifically in the tonsils of patient 2. A pair of tonsils removed at the same time because of benign hypertrophy served as the controls. No gross differences in the rates of sterol synthesis from either acetate or mevalonic acid were detected between these two tissues.2 Recent measurements of lysosomal cholesteryl ester acid hydrolase in the spleen of patient 5 have revealed no deficiency in the activity of this enzyme."

Morpholo of Upid Deposits in Tangier Disea A consistent pattern of involvement of tissues and organs is evident on review of the histopathologic findings on various biopsies of patients 1-12, 14-17 and the autopsy observations on patient 13 (Table 2). As indicated in detail in Table 2, foam cells have been observed in: tonsils, lymph nodes, thymus, bone marrow, liver, spleen, skin, submucosa of small and large intestine, gall bladder, ureters and renal pelves, pyelo- nephritic scars in kidneys, small nerves, areas of lung which had been the sites of bronchopneumonia, tunica albuginea of testicles and in mitral and tricuspid valves. The central nervous system was found to be uninvolved in the only patient (13) in which it was examined. Although the foam cells present in the tissues enumerated above have been considered to be lipid-laden histiocytes, the present study shows that Schwann cells and smooth muscle cells also contain numerous lipid deposits. The present study complements previous reports 2-14 showing that the lipid deposits in foam cells in various organs of the reported patients with Tangier disease are histochemically similar. These deposits are birefringent and are stainable by lipid-soluble dyes such as oil red 0, Sudan black B, Sudan III and Scharlach R, and by the Schultz reaction for cholesterol. The periodic acid-Schiff reaction for vicinal glycols, the Rinehart stain for acid mucopolysaccharides and the Baker stain for phospholipids are negative. These results are concordant with the extensive chemical data (see Fredrickson "5 for review) showing that an increased content of cholesteryl esters is the consistent abnormality. Analyses have specifically shown increased concentrations of cholesteryl esters in tonsils (patients 1, 2 and 3), spleen (patient 5), lymph node (patient 1), skin (patient 5) and rectal mucosa (patient 9). The con- 120 FERRANS AND FREDRICKSON American Journal of Pathology centrations of unesterified cholesterol in tonsils (patients 1, 2 and 3), spleen (patient 5) and lymph node (patient 1) were not increased. Elevated content of triglycerides was found in the spleen of patient 5, but not in the lymph node of patient 1. The ultrastructural data also are consistent with the concept that cholesteryl esters are the major component of the lipid deposits. The completely electron-lucent appearance of most of the areas of deposition in tonsils and jejunum indicates that their contents had been dis- solved during preparation for electron microscopy. The jagged outlines left by many such deposits and their crystal type of birefringence indicate that many of the lipid deposits are crystalline. Not all of the lipid is stored in solid crystals, however, for numerous droplets in bone marrow foam cells exhibited the Maltese cross type of birefringence, and some tonsillar deposits were not birefringent. The significance of these variations in birefringence is uncertain, particularly with respect to observations made in sections stained with oil red 0; it has been our experience that this stain can produce artifactual crystals. But some of the stored lipid could be triglycerides or other nonsterol material. For example, the material of low electron density which was retained in some of the lipid droplets in foam cells in jejunum, tonsils and bone marrow also may have been triglyceride. Furthermore, particles in the size range of chylomicrons were present within phagocytic vacuoles in some foam cells in bone marrow, and most foam cells in tonsils, jejunum and bone marrow also contained material with the characteristics of ceroid. Evi- dence reviewed in detail elsewhere 18 indicates that ceroid deposits represent the end stage of polymerization and peroxidation of lipids. From the preceding we conclude that the foam cells in Tangier disease contain the following four types of deposits: a) crystals of cholesteryl esters; b) droplets which may represent mixtures of cholesteryl esters and triglycerides; c) pleomorphic masses of ceroid, and d) particles in the size range (1,000 to 5,000 A) of chylomicrons.

Comparisons of Morphologic Changes in Tangier Disease and in Other Diseases The morphologic features of Tangier disease require comparisons with those of certain other diseases that fall into three broad categories. The first includes some of the disorders in which deficiency of a specific acid hydrolase leads to the accumulation of lipids within lysosomes or lysosome-derived structures. The second category comprises various other diseases in which tissue lipid storage results from metabolic abnormalities not related to deficiencies of lysosomal hydrolases. The third includes certain proliferative diseases of the reticuloendothelial Vol. 78, No. 1 TANGIER DISEASE 121 January 1975 system which sometimes are associated with cholesterol storage in abnormal histiocytes. Disorders which involve lysosomal hydrolase deficiencies and which must be distinguished from Tangier disease include the sphingolipidoses, cholesteryl ester storage disease, Wolman's disease, and the mucopolysaccharidoses and mucolipidoses. Among the sphingolipidoses, Niemann- Pick disease is associated with extensive storage of cholesterol. The sterol is not esterified and occurs in amounts almost equimolar to those of the sphingomyelin deposited because of sphingomyelinase deficiency.3" The lipids are deposited in foam cells which are present throughout nearly all tissues and which by light microscopy are indistinguish- able from those in Tangier disease. The deposits are organized in the form of electron-dense lamellae, surrounded by trilaminar membranes, and thus differ from the deposits in Tangier disease. In other generalized sphingolipid storage diseases, neither the staining characteristics nor the appearance of the deposits should be confused with those in Tangier disease. The storage cells in Gaucher's disease appear distinctive by light microscopy. By electron microscopy the stored glucosyl ceramide (cerebroside) is present in tubules.404' In Fabry's disease ceramide trihexoside accumulates in highly organized, concentric or parallel lamellae.42A4 The sulfatide deposits in metachromatic leukodystrophy form inclusions which vary from finely granular to complex lamellar forms.45 In generalized gangliosidosis the neuronal storage of ganglio- sides is in the form of electron-dense, parallel lamellae."4647 In this disease the liver, spleen and other visceral organs may be filled with foam cells that contain fine tubules, up to 200 A in diameter, presumably composed of mucopolysaccharides.47 Cholesteryl ester storage disease (CESD) and Wolmaji's disease are two biochemically related, but clinically different diseases featuring generalized storage of cholesteryl esters and triglycerides. They are due to deficiency of one or more lysosomal hydrolases which normally cat- alyze the splitting of these esters.2' Wolman's disease is further characterized by calcification of adrenal glands, severe hepatosplenomegaly, steatorrhea, and death in infancy. Patients with CESD follow a much more benign course and have marked hepatomegaly without symptoms and without evidence of hepatic dysfunction.2' The foamy histiocytes in Tangier disease are similar in light microscopic appearance and in certain histochemical characteristics to those found in CESD and in Wolman's disease, but they differ significantly in their ultrastructural features. The fundamental morphologic difference is that the lipid deposits in CESD 48.49 and in Wolman's disease m are hmited by wel defined 122 FERRANS AND FREDRICKSON American Journal of Pathology trilaminar membranes that are lysosomal in nature. We have observed that lysosomes and lysosome-derived residual bodies in foam cells in Tangier disease do contain crystal clefts and lipid droplets. Most of the accumulated lipids in this disease, however, are extralysosomal and not limited by true membranes. We have observed electron-dense lamellae, presumed to be composed of ceroid, at the edges of lipid droplets in lipid deposits in Tangier disease, Wolman's disease, CESD and in hyperlipoproteinemias. Similar lamellae also have been described in normal tissues.5' These lamellae should not be mistaken for true trilaminar membranes. Conversely, the trilaminar structure of true memberanes may not be evident when they are sectioned and viewed at oblique angles (rather than perpendicular to their surfaces) as is often the case in ultrathin sections of large lipid droplets. These sources of error must be considered carefully in the interpretation of electron micrographs of foam cells. Metabolic diseases due to defects other than deficiencies of lysosomal enzymes sometimes also are accompanied by the appearance of lipid- laden foam cells in bone marrow, liver, spleen and other tissues. These diseases include severe diabetes mellitus,52 type 1 (glucose-6-phospha- tase deficiency) glycogen storage disease 53 and certain of the "primary" hyperlipoproteinemias.8"95456 A rare example of the latter is familial lipoprotein-lipase deficiency (type 1 hyperlipoproteinemia), which is characterized by extreme hyperchylomicronemia and, apparently, secondary phagocytosis of circulating chylomicrons and VLDL by histiocytes.'819 The developing foam cells in type 1 hyperlipoproteinemia uindergo a series of morphologic changes related to the phagocytosis of chylomicrons and to subsequent metabolism of the lipids in these particles.19 In early stages of development, the foam cells in type 1 hyperlipoproteinemia and in Tangier disease are similar in the marked birefringence and presence of phagocytic vacuoles filled with particles measuring from 1,000 to 5,000 A in diameter, a size range comparable to that of chylomicrons. We have observed, however, that in severe hyperglyceridemic states the lipid deposits in foam cells evolve into large, intensely autofluorescent, electron-dense, nonbirefringent masses of ceroid.'8'"55"5 In contrast, the "mature'' foam cells in Tangier disease contain large, confluent, electron-dense, crystalline masses of cholesteryl esters and only small amounts of ceroid. Birefringent foam cells also occur in familial deficiency of lecithin-cholesterol acyltransferase (LCAT deficiency), a disorder in which lesser amounts of lipids, particularly unesterified cholesterol, accumulate in kidney and elsewhere.3157 These cells contain numerous vacuoles which often have Vol. 78, No. 1 TANGIER DISEASE 123 January 1975 osmiophilic peripheral zones. In addition, the spleen contains histiocvtes filled with concentrically arranged, electron-dense lamellae, and kidney tissues have masses of electron-dense membranes surrounding amorphous, mottled structures of electron-lucent spaces; these masses are present in capillary lumina, capillarv basement membranes, mesangial regions, capsular basement membranes and, occasionally, in the urinary space.57 Similar findings have not been reported in Tangier disease, and the clinical picture of chronic renal disease and anemia observed in LCAT deficiency has not been seen in Tangier disease. Histiocvtosis X is characterized by reticuloendothelial hyperplasia and by granuloma formation. The amount of histologically demonstrable lipid in the proliferating histiocvtes varies according to the stage of the disease, but the lesions often contain abnormal accumulations of lipids, particularly of cholesteryl esters.15'58 Ultrastructural studies have been made of the three variants of histocytosis X: Hand-Schiiller-Christian disease,59, Letterer-Siwe's disease 61I63 and eosinophilic granuloma.6u6 These studies have given little attention to the morphology of the lipid deposits but have disclosed a feature common to all three variants of histiocytosis X, namely, the consistent presence of Langerhans granules in the cvtoplasm of the histiocytes. Langerhans granules are pleomorphic, are limited bv single, trilaminar membranes, and varv from elongated rods to structures formed bv either one (tennis racquet image) or two rods attached to a vesicle. The rod-like portions of Langerhans granules contain a central stratum of electron-dense material which has a square lattice type of substructure with a bidirectional periodicity of 90 A.67 Langerhans granules are not specific for histo- cvtosis X. They are present also in normal Langerhans cells of the epidermis and have been found in abnormal histiocytes in various con- ditions other than histiocvtosis X (see Buja et al " for review). They are not present in histiocytes in Tangier disease or in other diseases of lipid metabolism. Sudanophilic lipid deposits also have been described in lymph nodes of patients with other rare histiocytic disorders variously described as sinus histiocytosis with massive lymphadenopathy,69 and chronic lymphadenopathy with reticulohistiocytic proliferation and lipid storage.70'7 In some of these patients, biochemical analyses of lymph nodes revealed the presence of increased amounts of cholesterol, cholesteryl esters and triglycerides.71 These disorders differ from histiocytosis X in that they follow a benign course and are not associated with other systemic manifestations. Studies of plasma lipids have been incomplete in patients with these syndromes of benign lymphadenopathy and lipid accumulation. The clinical picture and histology of these lesions (which 124 FERRANS AND FREDRICKSON American Journal of Pathology are mostly in cervical lymph nodes) should not be confused with those of Tangier disease.

Distribution of Lipid Deposits in Tangier Disease The distribution of lipid deposits in Tangier disease is of diagnostic importance because it differs in certain characteristic ways from that observed in other diseases of lipid metabolism. The tonsils character- istically show marked enlargement and a unique coloration which varies from yellow-orange to yellow-gray. These features have been noted in the tonsils of 9 patients (1-3, 11 and 13-17) and in tonsillar remnants from 6 others (patients 4-8 and 10). Histologic examination revealed numerous foam cells in the tonsils of 9 patients (1-3, 11 and 13-17) and in tonsillar remnants of patients 4, 6 and 8. No histologic studies of tonsils or tonsillar remnants were reported in the other patients. The distribution of the foamy histiocytes within bands of fibrous tissue surrounding the lymphoid follicles apparently accounts for the yellow stripes in the tonsils. Foam cells also have been reported in tonsils of patients with Gaucher's disease,72 Wolman's disease and Niemann-Pick disease,74 but in much lesser numbers than in Tangier disease and without remarkable changes in gross appearance. In the 2 patients (4 and 13) in whom the thymus was studied it was found to be infiltrated by foam cells to an extent comparable to that seen in tonsils. Clinically evident lymphadenopathy is not a feature of Tangier disease, but enlarged lymph nodes in 3 patients (cervical in 1, mesentric in 5 and cervical, mediastinal and abdominal in #13) were found to contain abundant foam cells. Infiltration of the thymus, spleen and lymph nodes by foam cells has also been reported in Nie- mann-Pick disease,74 Wolman's disease,73 cholesteryl ester storage disease,49 Fabry's disease,75 GM1 gangliosidosis 7¢ and Gaucher's disease.72 Some degree of enlargement of the spleen was present in 10 of the 17 patients with Tangier disease but was marked (1160 g) only in patient 5. The spleen was examined histologically in patients 5 and 13. Only a few foam cells were present in the spleen of patient 5, who was much older. Two additional observations on patient 5 are of particular interest. His spleen was removed because of leukopenia and thrombocytopenia.5 Soon after splenectomy he developed cutaneous lesions which measured from 2 to 3 mm in diameter and contained histiocytes filled with birefringent material. Chromatographic analysis of lipids extracted from these papules revealed the presence of large amounts of cholesteryl oleate. It is quite noteworthy that prior to splenectomy, patient 5 also had very low plasma cholesterol concentration (30-60 Vol. 78, No. 1 TANGIER DISEASE 125 January 1975 mg/100 ml). Eight months after splenectomy his plasma cholesterol concentration was noticeably higher (100-155 mg/100 ml), but without alteration in his very low levels of plasma HDL. These observations suggest that the spleen in this patient may have played an important role in the removal of lipids from plasma, possibly by the phagocvtic action of reticuloendothelial cells. The liver has been found to be mildly enlarged in 5 patients (1, 5, 6, 10 and 16), and foam cells have been observed in the livers of 4 patients (5, 6, 10 and 13). In each of these patients the foam cells were few in number, were identified as histiocytes, and showed the same histochemical characteristics of foam cells in other organs. No involvement of hepatic parenchymal cells was observed in patients 10 and 13, but the hepatic parenchymal cells in patients 5 and 6 contained small granules with the histochemical properties of ceroid deposits. This absent or minimal involvement of hepatocytes in Tangier disease is in sharp contrast to that in Wolman's disease and cholesteryl ester storage disease, both of which are associated with marked hepatomegaly and extensive lipid deposition in the liver. The color of the liver is butter-yellow or orange in CESD49'7778 and yellow in Wolman's disease.7379 Most of the hepatic storage of lipids in CESD occurs in the form of birefiingent crystals within hepatocytes; histiocytes in liver contain primarily ceroid.49 In Wolman's disease, crystal clefts are found only in Kupffer cells (which also contain ceroid), but the hepatocytes show marked steatosis.79'30 The hepatic lipid deposits in these 2 diseases are intra- lysosomal in location and are limited by trilaminar membranes.4840 Bone marrow preparations were examined in 13 patients. Foam cells were present in 9 of these patients ( 1, 2, 5-8, 10, 15 and 16). Foam cells have also been found in the bone marrow of the mother of patients 1 and 2 and of the father of patient 16; both were considered to be heterozygous for the gene involved in Tangier disease. Nevertheless, in three other obligate heterozygotes we have failed to find marrow foam cells. Our experience with examination of bone marrow in patients with various lipidoses has been that the foam cells are much more easily detected in fresh preparations examined by polarization, phase contrast and interference microscopy than in fixed smears stained according to Wright's method. Many of these foam cells are extremely fragile and easily disrupted during the preparation of smears, and the need for the use of these optical techniques in the study of bone marrow of patients with suspected lipidoses is emphasized by the fact that in patients 1, 2 and 15 foam cells were not found on routine examination of smears of bone marrow stained by Wtright's method. As demonstrated in this study, the 126 FERRANS AND FREDRICKSON American Journal of Pathology bone marrow foam cells from patient 16 were ultrastructurally similar to those from his father. Both the small and large intestine are sites of infiltration by foam cells in Tangier disease. Symptoms of malabsorption or of other types of intestinal dysfunction have been reported in only 2 patients with Tangier disease. One of these patients (9) also had ulcerative colitis and underwent colectomy; this patient had a flat oral glucose tolerance curve. The other patient (5) had evidence of minimal malabsorption of fat, and had mild, chronic diarrhea, which responded to a low-fat diet and has not recurred. Foam cells are very numerous in rectal mucosa, where they form plaque-like lesions that are yellow to orange-reddish in color and measure up to several mm in diameter. This abnormal appearance, quite striking upon sigmoidoscopy, was found in the 7 patients (5, 7-10, 12 and 13) in whom the rectal mucosa was examined. Grossly evident rectal lesions of the type described above have not been reported in other diseases of lipid metabolism, although areas of yellow coloration in colonic mucosa have been reported in Wolman's disease,73 Whipple's disease and other intestinal granulomatous diseases.81 Involvement of the small intestine in Tangier disease appears to be less extensive than that of the rectum, as shown by the necropsy findings in patient 13 and by observations on jejunal biopsies of patients 2, 5 and 6. Of the latter 3 patients, only patient 2 had foam cells in jejunum. The localization of foam cells in the submucosa of jejunum, deep to the muscularis mucosa, was similar in patients 2 and 13; in both, the mucosal villi were free of foam cells. These findings differ from those reported in CESD 48 and in Wolman's disease 73,79-82 in both of which diseases the small intestinal mucosa is yellow to orange in color and the villi may be infiltrated by foam cells. The villi in Tangier disease and in CESD 48 are normal in size, whereas in Wolman's disease 'I they are thick, flattened and club-shaped. Lipid deposits and foamy cytoplasmic vacuolization in smooth muscle cells of the muscularis mucosa have been reported in CESD.48 No ultrastructural studies of jejunal mucosa have been reported in Wolman's disease or in any of the familial hyperlipoproteinemias. In contrast to the findings in Tangier disease, the infiltration of foamy histiocytes in Wolman's disease,m CESD 48,49 and generalized gangliosidosis 83 iS more prominent in the small than in the large intestine. In the infantile form of Gaucher's disease, foam cells also are present in the lamina propria of the small intestine.83 The findings in Tangier disease also differ from those in abetalipoproteinemia (Bassen- Kornzweig syndrome), in which marked accumulation of lipid (mostly triglycerides) occurs in the cytoplasm of mucosal epithelial cells but not Vol. 78, No. 1 TANGIER DISEASE 127 January 1975 elsewhere in the mucosa.- Variable numbers of foamy histiocytes also are found in intestinal and rectal mucosa and submucosa in Niemann- Pick disease, Fabry's disease, Hurler's disease, metachromatic leukodystrophy and in some of the mucopolysaccharidoses.m383 As discussed previously, foam cells in these diseases can be distinguished from those in Tangier disease on the basis of their ultrastructural and cytochemical characteristics. Foamy histiocytes also can be found in rectal mucosa of patients with a variety of diseases unrelated to disorders of lipid metabolism;86 such histiocytes, particularly those in the syndrome of pig- mented colonic histiocytosis,87 usually are strongly PAS-positive and are filled with electron-dense bodies of pleomorphic ultrastructure. Thus, they differ from those in Tangier disease. Examination of skin was performed in 11 patients (1-6, 8-10, 13 and 16) and cutaneous deposits of cholesteryl esters were found in 7 (1, 3, 4, 5, 10, 13 and 16). Based on light microscopic studies, Waldorf et al1 concluded that many of the cholesteryl ester deposits in the skin of patients with Tangier disease were present in extracellular locations in the dermis. Such deposits were not observed in samples of clinically normal skin from 3 normal patients and from 5 patients with Niemann- Pick disease.4 Cutaneous xanthomas with tvpical foam cells rarely occur in Niemann-Pick disease,74 but are common in certain types of hyper- lipoproteinemiaI8'l and in biliary cirrhosis.89 Only one patient with Tangier disease (5) has had visible cutaneous lesions, and these have been confined to small papular lesions. Extracellular deposits of cholesterol and cholesteryl esters have been described in the skin of patients with a rare syndrome known as extracellular cholesterolosis of Urbach (see Waldorf et a14 for review). The skin lesions in extracellular cholesterolosis are large (up to several cm in diameter), often are ulcerated, show marked vasculitis, and are infiltrated by neutrophils, lymphocytes, histiocytes and occasionally also by eosinophils and plasma cells.4 These features are not observed in Tangier disease. Extracellular cholesterolosis is now considered to be a variant of erythema elevatum diutinum (a bizarre, uncommon type of cutaneous vasculitis of unknown etiology) with secondary lipid deposits.4 Our observations on skin biopsies from 3 patients (4, 5 and 16) show that clinically normal-appearing skin is the site of lipid deposits in Tangier disease, and that these deposits are localized to dermal histiocytes and to Schwann cells of small cutaneous nerves. We did not find any evidence of extracellular deposits of cholesteryl esters in either of these 3 patients. Involvement of Schwann cells appeared to be quanti- tatively more important than that of macrophages. The small size of 128 FERRANS AND FREDRICKSON American Journal of Pathology many affected nerves probably accounts for the apparent extracellular localization of the lipid deposits observed by light microscopy. Neurologic abnormalities have been detected clinically in 10 patients (1, 2, 5, 7-10, 12, 13 and 15) with Tangier disease. One patient (13) had severe mental retardation and spasticity which were attributed to cerebral trauma at the time of birth. Neurologic manifestations in the other 9 patients have been those of peripheral neuropathy, which is asymmetric in distribution and shows marked fluctuations from time to time, with even complete remissions. These manifestations have included: a) muscular weakness in either the upper or lower extremities, with denervation type of -muscular atrophy; b) sensory loss; c) pares- thesias; d) diminished or absent reflexes, and e.) disturbances of external ocular movements. Anatomic studies of nervous tissue have been made in 9 patients with Tangier disease, including observations made by light microscopy in 3 patients (8, 9 and 13) and by both light and electron microscopy in 5 others (2, 4, 5, 7, 12 and 16). No lipid storage was found either in neurons or macrophages in the central nervous system of the only necrop- sied patient (13). The appearance of the myenteric plexus was not described in this patient; however, a few pale macrophages were observed around small nerves in the skin, but not around larger subcutane- ous nerves. In contrast to this, light microscopic study of nerve biopsies in patients 8 and 9, both of whom had peripheral neuropathy, did not show any lipid storage. In radial nerve (patient 7) and sural nerve (patient 12) biopsies Kocen et al 7 found the following ultrastructural changes: a) loss of unmyelinated and myelinated axons; b) lipid deposits (mostly electron-lucent, but also a few composed of electron- dense lamellae) in the cytoplasm of Schwann cells, and c) endoneurial fibrosis. The only change found in vascular elements in nerves was reduplication of the basement membranes. We did not study large peripheral nerves, but in small nerves in clinically normal skin (patients 4, 5 and 16) and small intestine (patient 2) we found extensive lipid deposits in practically every Schwann cell and variable degrees of endoneurial fibrosis. Lipid deposits in Schwann cells are not specific for Tangier disease. Accumulation of lipids in Schwann cells and neurons of myenteric plexus have been reported in Tay-Sachs disease, the G5 variant of systemic late infantile lipidosis, the Jansky-Bielchowsky type of late infantile amaurotic idiocy and Niemann-Pick disease.85 Lipid deposits in Schwann cells of peripheral nerves have been found in metachromatic leukodystrophy,45 Krabbe's disease,90 Fabry's disease 42,43 and in Refsum's dis- Vol. 78, No. 1 TANGIER DISEASE 129 January 1975 ease.91 In Tay-Sachs disease, the G5 variant of systemic late infantile lipidosis, the Jansky-Bielchowsky type of late infantile amaurotic idiocy and Niemann-Pick disease,83 as well as in Krabbe's disease," the lipid deposits are more extensive in neurons than in Schwann cells. In contrast to this, neurons and axons are not involved in CESD, but Schwann cells are.48 Light microscopic studies have revealed neuronal lipid deposits in myentric plexus in Wolman's disease;m the ultrastructure of Schwann cells or axons in this disease has not been described. Among the diseases of lipid metabolism, demyelination and increased amounts of endoneurial collagen in peripheral nerves have been reported in Krabbe's disease," in abetalipoproteinemi,92 and in 1 patient (7) with Tangier disease. The peripheral nerve lesions in Refsum's disease are further characterized by a hypertrophic neuropathy, with onion bulb- like whorls, and by crystalline inclusions in mitochondria of Schwann cells.91 In biliary cirrhosis associated with severe hypercholesterolemia, cutaneous xanthomata and mild sensory neuropathy, Kocen et al found massive collections of lipid-laden macrophages in peripheral nerves, and lipid-containing vacuoles in perineurial cells but not in Schwann cells.7

Theories of Pathogenesis of lpid Strwage in Tangrer Disease The basis for the intracellular accumulation of cholesteryl esters in Tangier disease is still almost entirely speculative. Lacking is certain essential knowledge of the physiologic roles of HDL and of many aspects of the metabolism of other lipoproteins and cholesteryl esters. The application of available information to Tangier disease is also incomplete, the rarity of the patients and inaccessibility of certain tissues having left certain avenues only partially explored. Although the data obtained thus far suggest neither uncontrolled cholesterol synthesis nor deficient ability to hydrolyze sterol esters in Tangier disease, it is desir- able that these processes be examined more thoroughly. Also lacking is information concerning the localization of apoproteins within the storage cells, the effective levels of lipoprotein lipase and LCAT activity, as opposed to their in vitro measurements, and the plasma turnover rates of the separate apoproteins. The augmentation in tissue concentration of only a single class of lipids does not exclude a defect involving other lipids or lipoproteins. An analogous situation may exist in atheromatous plaques, which are believed to arise partly from uptake of lipoprotein complexes, yet cholesteryl esters are the most prominent lipid components of plaques. Choles- teryl esters, especially if they become crystalline, are less soluble than phospholipids and presumably less susceptible to enzymatic attack or 130 FERRANS AND FREDRICKSON American Journal of Pathology other removal processes. Therefore, they may accumulate selectively, even though they may have originally arrived in the company of other lipids. If we exclude a primary defect in synthesis of cholesterol or enzymatic degradation of its esters, the pathogenesis of Tangier disease mainly revolves around two major possible abnormalities. One is the uptake of lipid by a greatly enhanced process of phagocytosis or pinocytosis. The other is the failure of some process that normally removes cholesterol synthesized locally in cells to some other site. The available biochemical information provides support, but not proof, of the first of these mechanisms. The presence in bone marrow of some foam cells with membrane-bounded inclusions is in accord with uptake by endocytosis of fat particles. These could be chylomicrons, VLDL or even LDL, rendered abnormal in composition and stability by the absence of HDL. A deficiency in C-proteins or possible interference with the enzymatic processes normally believed to be important in triglyceride metabolism has been mentioned, but some more critical role for HDL in triglyceride transport may yet be discovered. It is also possible that the aggregates of lipid found in lysosomes are remnants of unstable HDL, secreted in significant quantities but incapable of surviv- ing in circulation. The theory that lipid deposition is due to collection of abnormal lipoproteins or metabolic remnants by phagocytosis as the initial basis for lipid deposition is supported by indirect evidence. Abnormal lipoproteins particles have been isolated from Tangier plasma.26 Secondly, the kinds of cells involved suggest uptake of "foreign bodies." Microscop- ically, lipid droplets are mainly restricted to tissues rich in reticuloendothelial cells, although nonvascular smooth muscle cells and particularly Schwann cells seem to be important exceptions. It is noteworthy that endothelium of small blood vessels or capillaries is not vacuolated. The already noted influence of splenectomy on plasma cholesterol levels of patient 5 supports the concept that phagocytosis plays a major role in the pathogenesis of the tissue storage of cholesteryl esters. If uptake of lipoproteins by phagocytosis is the principal initiating event, the scavenged material must undergo rapid transformations. Lipids other than cholesterol are disposed of, since no evidence of their accumulation is obtained upon gross analyses of involved tissues. More- over, any initial limiting membrane is disposed of, for the vast majority of cholesteryl ester deposits lie unbounded within the cytoplasm. It is this latter feature which leads to speculation of whether all of the excess steryl esters are, in fact, derived from the scavenging of aberrant Vol. 78, No. 1 TANGIER DISEASE 131 January 1975 lipoproteins. It is possible that cholesterol or cholesteryl esters also are imbibed from entracellular fluid without involving membranous seques- tration. Regardless, there is no effective process for removal of this excess and it continues to accumulate. It remains for future work to determine whether HDL performs some role in sweeping excess sterol from cells into extracellular fluid, thus maintaining normal intracellular concentrations. The localization of lipid deposition in Tangier disease also raises questions that cannot be answered. First among these is the extraordinary involvement of the tonsils; another is the restriction of foam cells in the small intestine to the submucosa. Is the lack of abnormal storage in the jejunal mucosa due to removal of lipids from this region by their incorporation into chylomicrons and other lipoproteins? Finally, there is apparently extraordinary involvement of the Schwann cells. The appearance of lipid droplets that distend the cytoplasm of these cells suggests that they interfere with the axons they surround, possibly through compression or physical distortion. Presumably, the relapsing nature of the polyneuropathy in Tangier disease may be due to cyclical changes in the degree of lipid accumulation within the Schwann cells. Eventually, however, the involvement does lead to axonal degeneration and loss. Tangier disease represents an unusual opportunity to determine the pleomorphic effects which so often attend mutational change in a single protein. Sometimes such disorders also make clear the normal function of the molecule or system most directly affected. In many ways Tangier disease has provided valuable information about the degree to which HDL are important for life and for certain normal functions. A full explanation of why we have HDL-or tonsils-must be derived from further studies of both this disease and the normal physiology of fat transport and metabolism.

References 1. Fredrickson DS, Gotto ANM, Levy, RI: Familial lipoprotein deficiency (abetalipoproteinemia, hypobetaipoproteinemia, Tangier disease), The Met- abolic Basis of Inherited Disease, Third edition. Edited by JB Stanbury, JB Wyngaarden, DS Fredrickson. New York, McGraw-Hill Company, 1972, pp 493-530 2. Fredrickson DS, Altrocchi PH, Avioli LV, Goodman DS, Goodman HC: Tangier Disease. Combined Clinical Staff Conference at the National Institutes of Health. Ann Intern Med 55:1016-1031, 1961 3. Fredrickson DS, Atrocchi PH: Tangier disease (familial cholesterolosis with high density lipoprotein deficiency), Cerebral Sphingolipidoses: A Sym- 132 FERRANS AND FREDRICKSON American Journal of Pathology

posium on Tay-Sachs' Disease and Allied Disorders. Edited by SM Aronson, BW Volk. New York, Academic Press, 1962, pp 343-357 4. Waldorf DS, Levy RI, Fredrickson DS: Cutaneous cholesterol ester deposition in Tangier disease. Arch Dermatol 95:161-165, 1967 5. Hoffman HN, Fredrickson DS: Tangier disease (familial HDL deficiency): clinical and genetic features in two adults. Am J Med 39:582-593, 1965 6. Kocen RS, Lloyd JK, Lascelles PT, Fosbrook AS, Williams D: Familial a- lipoprotein deficiency (Tangier disease) with neurological abnormalities. Lancet 1:1341-1345, 1967 7. Kocen RS, King RHM, Thomas PK, Haas LF: Nerve biopsy findings in two cases of Tangier disease. Acta Neuropathol 26:317-327, 1973 8. Engel WK, Dorman JD, Levy RI, Fredrickson DS: Neuropathy in Tangier disease: a-lipoprotein deficiency manifesting as familial recurrent neuropathy and intestinal lipid storage. Arch Neurol 17:1-9, 1967 9. Kummer H, Laissue J, Spiess H, Pflugshaupt R, Bucher U: Familiare Analphalipoproteinamie (Tangier-Krankheit). Schweiz Med Wochenschr 98:406-412, 1968 10. Kracht J, Huth K, Schoenborn W, Fuhrmann W: Hypo-a-Lipoproteinamine (Tangier Disease). Verh Dtsch Ges Pathol 54:355-360, 1970 11. Haas LF, Bergin JD: Alpha lipoprotein deficiency with neurological features. Aust Ann Med 1:76, 1970 12. Bale PM, Clifton-Bligh P, Benjamin BN, Whyte HM: Pathology of Tangier disease. J Clin Pathol 24:609-616, 1971 13. Clifton-Bligh P, Nestel PJ, Whyte HM: Tangier disease: report of a case and studies of lipid metabolism. N Engl J Med 286:567-571, 1972 14. Greten H, Hannemann T, Gusek W, Vivell 0: Studies on lipoproteins and lipolytic plasma enzymes in a new case of Tangier disease. (In press) 15. Fredrickson DS: Familial high-density lipoprotein deficiency: Tangier disease, The Metabolic Basis of Inherited Disease, Second edition. Edited by JB Stanbury, JB Wyngaarden, DS Fredrickson. New York, McGraw-Hill Book Company, 1966, pp 486-508 16. Fredrickson DS, Levy RI, Lees RS: Fat transport in lipoproteins: an in- tegrated approach to mechanisms and disorders. N Engl J Med 276:34-42, 94-103, 148-156, 215-225, 273-281, 1967 17. Freeman JA, Spurlock BO: A new epoxy embedment for electron microscopy. J Cell Biol 13:437-443, 1962 18. Ferrans VJ, Buja LM, Roberts WC, Fredrickson DS: The spleen in type 1 hyperlipoproteinemia: histochemical, biochemical, microfluorimetric and electron microscopic observations. Am J Pathol 64:67-96, 1971 19. Ferrans VJ, Roberts WC, Levy RI, Fredrickson DS: Chylomicrons and the formation of foam cells in type 1 hyperlipoproteinemia: a morphologic study. Am J Pathol 70:253-272, 1973 20. Ferrans VJ, Fredrickson DS: Unpublished data 21. Sloan HR, Fredrickson DS: Rare familial diseases with neutral lipid storage Wolman's disease, cholesteryl ester storage disease and cerebrotendinous xanthomatosis.1 22. Assmann G, Fredrickson DS, Sloan HR, Fales HM, Highet Rj: Accumula- tion of oxygenated steryl esters in Wolman's disease. J Lipid Res (In press) 23. Menkes JH, Schimschock JR, Swanson PD: Cerebrotendinous xanthomatosis: Vol. 78, No. 1 TANGIER DISEASE 133 January 1975

the storage of cholestanol within the nervous system. Arch Neurol 19:47-53, 1968 24. Bhattacharyya AK, Connor WE: ,B-sitosterolemia and zanthomatosis: a newly described lipid storage disease in two sisters. J Clin Invest 53:1033- 1043, 1974 25. Fredrickson DS, Levy RI, Lindgren FT: A comparison of heritable abnormal lipoprotein patterns as defined by two different techniques. J Clin Invest 47:2446-2457, 1969 26. Asmann G, Fredrickson DS, Herbert PN, Forte GM, Heinen RJ: An A-II lipoprotein particle in Tangier diseas. Ciculation (In press) 27. Herbert PN, Fredrickson DS, Heinen RJ, Forte GM, Assmann G: Tangier disease: a pathogenetic nism. Circulation (In press) 28. Heinen RJ, Herbert PN, Fredrickson DS: High density lipoproteins as a reservoir for C-apolipoproteins. Circulation (In press) 29. Lux SE, Levy RI, Gotto AM, Fredridcson DS: Studies on the protein defect in Tangier disease: isolation and characterization of an abnormal high density lipoprotein. J Clin Invest 51:2505-2519, 1972 30. Fredrickson DS: Plasma lpoproteins and apolpoproteins, Harvey Lectues. New York, Academic Press, 1974 (In press) 31. Glomset JA, Nonrm KR: The metabolic role of lecithin: cholesterol acyltransferase: perspectives from pathology. Adv lipid Res 11:1-65, 1973 32. Kostner G, Holasek A: Characterization and quantitation of the apolipoproteins from human chyle chylomicrons. Biochemistry 11:1217-1223, 1972 33. Barker WC, Dayhoff MO: Computer studies of distantly related sequences. Biophysical Society Abstracts, Seventeenth Annual Meeting, Feb 27-Mar 2, 1973, p 205a 34. Assmann G, Brewer HB Jr: lipid-protein interactions in high density lipoproteins. Proc Nail Acad Sci USA 71:989-993, 1974 35. Assmann G, Highet RJ, Sokoloski EA, Brewer HB Jr: 13C-nuclear magnetic resonance spectrscopy of native and recombined lipoproteins. Proc Nat] Acad Sci (In press) 36. Assmann G, Brewer HB Jr: A molecular model of high density lipoproteins. Proc Nail Acad Sci USA 71:1534-1538, 1974 37. Krauss R, Fredrickson DS: Unpublished data 38. Demosky, S, Sloan HR, Fredrickson DS: Unpublished data 39. Fredrickson DS, Sloan HR: Sphingomyelin lipidoses: Niemann-Pick disease.1 pp 783-807 40. Lee RE, Balcerzak SP, Westerman MP: Gauchers disease: a morphologic study and measurements of iron metabolism. Am J Med 42: 891-898, 1967 41. Hibbs RG, Ferrans VJ, Cipriano PR, Tardiff KJ: A histocheniical and eec- tron microscopic study of Gaucher cefls. Arch Pathol 89:137-153, 1970 42. Frost P, Tanaka Y, Spaeth GL: Fabry's disease-glycolipid lipidosis: histochemical and eletron microscopic studies of two cases. Am J Med 40:618- 627, 1966 43. Bischoff A, Fierz U, Regli G, Ulrich J: Peripher-neurologische Stirungen bei der Fabryschen Krankheit (Angiokeratoma corporis diffusum universale): klinisch-elektronenmikroskopische Befunde bei einem Fall. KIin Wochenschr 46:666-671, 1968 134 FERRANS AND FREDRICKSON American Journal of Pathology

44. Ferrans VJ, Hibbs RG, Burda CD: The heart in Fabry's disease: a histochemical and electron microscopic study. Am J Cardiol 24:95-110, 1969 45. Anzil AP, Blinzinger K, Martinius J: Ultrastructure of storage materials in metachromatic leucodystrophy peripheral neuropathy and in rabbit tissues after sulfatide application with special reference to pleated lamellar systems. J Microscopie 18:173-186, 1973 46. Gonatas NK, Gonatas J: Ultrastructural and biochemical observations on a case of systemic late infantile lipidosis and its relationship to Tay-Sachs disease and gargoylism. J Neuropathol Exp Neurol 24:318-340, 1965 47. Suzuki K, Suzuki K, Chen GC: Morphological, histochemical and biochemical studies on a case of systemic late infantile lipidosis (generalized gangliosidosis). J Neuropathol Exp Neurol 27:15-38, 1968 48. Partin JC, Schubert WK: Small intestinal mucosa in cholesterol ester storage disease: a light and electron microscopic study. Gastroenterology 57:542-.558, 1969 49. Frederickson DS, Sloan HR, Ferrans Vj, Demosky SJ: Cholesteryl ester storage disease: a most unusual manifestation of deficiency of two lysQsomal enzyme activities. Trans Assoc Amer Phys 85:109-119, 1972 50. Lake BD, Patrick AD: Wolman's disease: deficiency of E600-resistant acid esterase activity with storage of lipids in lysosomes. J Pediat 76:262-266, 1970 51. Fawcett DW: The Cell: Its Organelles and Inclusions: An Atlas of Fine Structure. Philadelphia, WB Saunders, 1966, p 311 52. Warren S, Root HF: Lipoid-containing cells in the spleen in diabetes with lipemia. Am J Pathol 2:69-80, 1926 53. Fleischmajer R: The Dyslipidoses. Springfield, Illinois, Charles C Thomas, Publisher, 1960, pp 178-187 54. Tanaka Y, Brecher G, Fredrickson DS: Cellules de la maladie de Niemann- Pick et de quelques autres lipoidoses. Nouv Rev Franc Hematol 3:5-16, 1963 55. Roberts WC, Levy RI, Frederickson DS: Hyperlipoproteinemia: a review of the five types with first report of necropsy findings in type III. Arch Pathol 90:46-56, 1970 56. Roberts WC, Ferrans VJ, Levy RI, Fredrickson DS: Cardiovascular pathology in hyperlipoproteinemia: anatomic observations in 42 necropsy patients with normal or abnormal serum lipoprotein patterns. Am J Cardiol 31:557- 570, 1973 57. Hovig T, Gjone E: Ultrastructural aspects of familial lecithin-cholesterol acyltransferase deficiency. Nutr Metab 15:89-96, 1973 58. Thannhauser SJ: Lipidoses. Diseases of the Intracellular Lipid Metabolism, Third Edition. New York, Grune & Stratton, 1958, pp 345-432 59. Gianotti F, Caputo R: Skin ultrastructure in Hand-Schuller-Christian disease: report on abnormal Langerhans' cells. Arch Dermatol 100:342-349, 1969 60. Imamura M, Muroya K: Lymph node ultrastructure in Hand-Schuller- Christian disease. Cancer 27:956-964, 1971 61. Cancilla PA, Lahey ME, Carnes WH: Cutaneous lesions of Letterer-Siwe disease: electron microscopic study. Cancer 20:1986-1991, 1967 62. Gianotti F, Caputo R, Ranzi T: Ultrastructural study of giant cells and "Langerhans cell granules" in cutaneous lesions and lymph-node and liver Vol. 78, No. 1 TANGIER DISEASE 135 January 1975

biopsies from four cases of subacute disseminated histiocytosis of Letterer- Siwe. Arch Kui Exp Dermatol 233:238-252, 1958 63. Kondo Y, Toioka H, Itoh M: Letterer-Siwe diseas and esinopbilic gnu- loma: case report including the ultartural findings of the bistiocyts prolifeating in these disorders. Acta Pathol Jap 19:525-536, 1969 64. Basset F, Nizelof C, Mallet R, Turiaf J: NouveIle mise en 6vidence, par la microscopie electronique, de particules d'allure virale dans une seconde forme clinique de Thistiocytose X, le granulome 6osinophile de ros. Compt Rend Acad Sci (Paris) 261:5719-5720, 1965 65. Basset F, N&elof C, Turiaf J: Prisence en microscopie eltronique de structures filamenteuses orginals dans les lesions pulmonaires et osseuses de rhistiocytose X: stat actuel de la question. Bull Soc Med Hop Paris 117: 413-426, 1966 66. Morales AR, Fine G, Hom RC Jr, Watson JHL: Langerhans cells in a walized lesion of the eosinophilic granuloma type. Lab Invest 20:412-423, 1969 67. Tusques J, Pradal G: Analyse tridimensionelle des inclusions rencontr6es dans les histiocytes de lhistiocytose "X", en microscopie 6lectronique: con- paraison avec les inchlsions des celDules de Langerhans. J Microscopie 8:113- 122, 1969 68. Buja LM, Ferrans VJ, Maron BJ: Intracytoplasmic junctions in cardiac muscle celkl. Am J Pathol 74:613-648, 1974 69. Rosai J, Dorfman RF: Sinus histiocytosis with massive lymphadenopathy: a newly recognized benign clinicopathological entity. Arch Pathol 87:63-70, 1969 70. Destombes P: Adenites avec surcharge lipidique, de renfant ou de radulte jeune, observ&s aux Antilles et au Mali (Quatre observations). Bull Soc Pathol Exot 58:1169-1175, 1965 71. Marie J, Bernard J, N6zelof C, Lkv6que B, Schaison G, Desbois JC, Watchi JM, Lemaigre-Voreaux J: Adenopathies chroniques avec prolifration r6ticu- lohistiocytaire et surcharge lipidique. Ann Pediatr 13:2689-2697, 1966 72. Reich C, Siefe M, Kessler BJ: Gaucher's disease: a review, and discussion of twenty cases. Medicine 30:1-20, 1951 73. Guanzi GC, Martin JJ, Philippart M, Roels H, van der Eecken H, Vrints L, Delbeke MJ, Hooft C: Wolman's disease. Eur Neurol 1:334-62, 1968 74. Crocker AC, Farber S: Niemann-Pick disease: a review of eighteen patients. Medicine 37:1-95, 1958. 75. Fessas P, Wintrobe MM, Cartwright GE: Angiokeratoma corporis diffusum universale (Fabry): first American report of a rare disease. Arch Intem Med 95:469-481, 1955 76. Landing BH, Silverman FN, Craig JM, Jacoby MD, Lahey ME, Chadwick DL: Familial neurovisceral lipidosis: an analysis of eight cases of a syndrome previously reported as "Hurler-Variant", 'pseudo-Hurler disease", and "Tay- Sachs disease with visceral involvement". Am J Dis Child 108:503-522, 1964 77. Lageron A, Caroli J, Stralin H, Barbier P: Polycorie chol1estroique de radulte. L Etude clinique, electronique, histochimique. Presse Med 75:2785- 2790, 1967 78. Schiff L, Schubert WK, McAdams AJ, Spiegel EL, O'Donnell JF: Hepatic cholesterol ester storage disease, a familial disorder. I. Clinical aspects. Am J Med 44:538-546, 1968 136 FERRANS AND FREDRICKSON American Journal of Pathology 79. Lough J, Fawcett J, Wiegensberg B: Wolman's disease: an electron microscopic, histochemical, and biochemical study. Arch Pathol 89:103-110, 1970 80. Lowden JA, Barson AJ, Wentworth P: Wolman's disease: a microscopic and biochemical study showing accumulation of ceroid and esterified cholesteroL Can Med Assoc J 102:402-405, 1970 81. Raffensperger EC: Chronic intestinal granulomas, Gastroenterology, Second edition, Vol II. Edited by HL Bockus. Philadelphia, W. B. Saunders, 1964, pp 404-419 82. Kyriakides EC, Filippone N, Paul B, Grattan W, Balint JA: Lipid studies in Wolman's disease. Pediatrics 46:431-436, 1970 83. Kamoshita S, Landing BH: Distribution of lesions in myenteric plexus and gastrointestinal mucosa in lipidoses and other neurologic disorders of children. Am J Clin Pathol 49:312-318, 1968 84. Dobbins WO: An ultrastructural study of the intestinal mucosa in congenital ,8-lipoprotein deficiency with particular emphasis upon the intestinal absorptive cell. Gastroenterology 50:195-210, 1966 85. Adachi M, Volk BW, Schneck L, Torii J: Fine structure of the myenteric plexus in various lipidoses. Arch Pathol 87:228-241, 1969 86. Yunis E, Sherman FE: Macrophages of the rectal lamina propria in children. Am J Clin Pathol 53:580-591, 1970 87. Pittman FE, Smith WT, Mizrahi A, Blanc WA, Pittman JC: Clinical, histochemical and electron microscopic study of colonic histiocytosis. Gut 7:458- 467, 1966 88. Bulkley BH, Buja LM, Ferrans VJ, Bulkley GB, Roberts WC: Tuberous xanthoma in homozygous type II hyperlipoproteinemia: a histologic, histochemical and electron microscopic study. Arch Pathol (In press) 89. Thomas PK, Walker JG: Xanthomatous neuropathy in primary biliary cirrhosis. Brain 88:1079-1088, 1965 90. Yunis EJ, Lee RE: Further observations on the fine structure of globoid leukodystrophy. Peripheral neuropathy and optic nerve involvement. Hum Pathol 3:371-388, 1972 91. Fardeau M, Engel WK: Ultrastructural study of a peripheral nerve biopsy in Refsum's disease. J Neuropathol Exp Neurol 28:278-294, 1969 92. Sobrevilla LA, Goodman ML, Kane CA: Demyelinating central nervous system disease, macular atrophy and acanthocytosis (Bassen-Kornzweig syndrome). Am J Med 37:821-828, 1964

Figures 1-18 are light micrographs of preparations made as indicated in the legend of each figure. Figures 19-40 are electron micrographs of sections stained with lead citrate and uranyl acetate. Figures 1-4 are sections of jejunal biopsy from patient 2. Fig 1-Numerous foam cells are subjacent to the muscular mucosa. The mucosa is not involved (H & E, x 75). Fig 2-Higher magnification view of part of section shown in Figure 1. Smooth muscle cells at top are vacuolated; foamy histiocytes surround small, thin-walled vessel. Compare with Figures 27 and 28, which show electron micrographs of similar areas (H & E, X 190). Figs 3 and 4-Two view of section stained with oil red 0 and examined by bright field (3) and polarization (4) microscopy. Foam cells are intensely stained and show marked birefringence (x 250). 4 ~ ~ ~~ ~ ~ ~~~~~~~~~~4

3 4 5 }

iW ' 1~~~~~~~~~~~~~~4.L MA ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,.!',ie;.A4t:iezjfN te,~~~4 ~ ' P 6 7

Fig 5-Group of foamy histiocytes in jejunal submucosa of patient 2, showing numerous cytoplasmic vacuoles of various sizes, and indentations in the nuclei. Compare with Fig- ures 19-21. Schwann cell (arrowheads) in small nerve at lower right also is vacuolated. Compare with Figure 29. Semithin (1 ,u) section of Maraglas-embedded tissue (Alkaline toluidine blue stain, X 1000). Fig 6Frozen section of tonsil stained with oil red 0 demonstrates preferential distribution of foam cells in septa between Iymphoid follicles. Part of a crypt is shown in center. Patient 2 (X 6). Fig 7-Large numbers of foamy histiocytes are present in septum between two Iymphoid follicles. Tonsil of patient 2 (H&E, X 100). 8 9

I 1 1 N 16 kw

Fgs 8 and 9-Two views of same area of oil-red-0-stained, frozen section of tonsil of patient 2. as viewed by bright field (8) and polarizaton (9) microscopy. The correspond- ence between the lipid staining and the birefringence is not complete (X 250). Fig 10- Large histiocytes with foamy, vacuobated cytoplasm and poorly defined cell borders are present in tonsil of patient 15. Compare with Figure 31, which shows uftrastructure of similar area. These cells are much larger than those in jejunum and in bone marrow (see Figures 5 and 11). Semithin section of Maraglas-embedded tissue (Alkaline toluidine blue stain, X 1000). Fig 11-Section of bone marrow of patent 15 showing several histiocytes with vacuolated cytoplasm. Semithin secton of Maraglas-embedded tissue (Alkaline toluidine blue stain, X 1000). ,#,. J.'

.1 ;

.1, i .,. t F-7

12 13

14 15

Figs 12-15-Four views of a foam cell in bone marrow of patient 15. This cell is shown as seen in an unstained fresh spread examined by bright field (12), polarization (13), phase contrast (14) and Nomarski interference contrast (15) microscopy. The cytoplasm of this cell is filled with refractile lipid deposits that show a Maltese cross type of birefringence (X 2000). 1 17 7~~~~~~~~~~~~~~~~~~~~1VIjl

Fig 16-Section of skin from patient 5 shows lipid deposits in a few scattered histiocytes (H) and in Schwann cells (SC) in a small nerve. Semithin section of Maraglas-embedded tissue (Alkaline toluidine blue stain, X 500). Figs 17 and 18-Views of two semithin sections through different levels of a small nerve in skin from patient 4. Marked degree of lipid accumulation is evident in the Schwann cells. Compare with Figure 39, which shows an electronmicrograph of this same nerve. Both sections were stained with alkaline toluidine blue. Section in 17 was photographed with ordinary bright field optics; section in 18 with Nomarski interference contrast optics (17, x 1200; 18, x 1600). W -w b 3 7 Fi§..f -

by a few collagen fibrls and microfibrils. They contain numerous lipid deposits, a few of which are partially coalescent, and some residual bodies. Most lipid de-posits are electron- lucent; others have moderately electron-dense rims. The nucleus of the cell in the center is indented by lipid deposits, and the cell surfaces have finger-like projections extending into the interstitium (x 7800). p t Y

F. :, O.e ;.

Lo.S*

"'0

__I

._Sei e 4

%6.

Fig 20-Two developing fbamy histocytes in jejunal submucosa of patient 2 contain a few lipid deposits, m ate numbers of mitochondria, numerous ribosomes and several Golgi cisterns. Compare with Figures 19 and 21 (x 15,000). 9.

I..

.1: L..1

Fig 21-Foamy histiocyte in jejunal submucosa of patient 2 is almost completely filled with large, coalescent lipid deposits which are separated from each other by thin strands of cytoplasm. The shapes of the lipid deposits are angular or irregular, rather than round; these shapes indicate the crystalline nature of the deposits. This cell contains few organelles and a few residual bodies and is surrounded by masses of collagen fibrils (X 6500). 4 -, 7 . .. 'r t, -,& '. 1, L04-1 I

4. .', 1. j

- lo..

-c

.r .,, .. II f

ia% .eC'. tp 4;. ---. "- -- 4F.; --.. .*,t .4op "A , 4w 1% qk

-- k A .0. ;

Fig 22-Foam cell contains lipid deposits, glycogen granules, 100-A-thick filaments and residual bodies. Crystal clefts of various shapes are seen in some of the residual bodies. Small Golgi cistems and a centriole are shown at lower left Jejunum of patient 2 (x 33,000). Figs 23-26-High magnification views showing the substructure of the edges of lipid deposits in foam cells. 23-Part of lipid deposit composed of clear center, a gray peripheral zone, and a single, electron-dense layer that does not have a trilaminar structure are shown. Note trilaminar plasma membrane (arrowheads). 24-shows lack of limiting membrane in lipid deposit close to the plasma membranes (arrowheads), which appear convoluted and and clearly trilaminar, is seen. 25-thin, single, electron-dense layers separating narrow strand of cytoplasm from two lipid deposits; note residual bodies (top and bottom) which contain electron-dense lipid material. 26-several parallel, electron-dense layers between 2 lipid deposits. Jejunum of patient 2 (23, x 93,000; 24, x 100,000; 25, X 76,500; 26, x 85,000). 23

24 25, 26 ,.f e -'.' -u.pJ

N,Wi

o 5 w .- x. !.. A.*

4 1, wi44 . t t s e, io tl*WI g: kffiffi.Y fi Fig 27-Low magnification view of arteriole in jejunal submucosa of patient 2. Foamy histiocytes (top and bottom left) surround the vessel. No lipid deposits are present in vascular smooth muscle or endothelium, although small lucent area (arrowheads) subjacent to the liendothelium'.S.'Sf4*may represent,<.f,feextracellular/ lipid (X 9600). J~~~~~~~~~~~

.e '4* >S . * t I 9

1. k) t' il Ak

)I I

ite't

FM!

v t9-1 ..A

Fig 28-Smooth muscle cell in muscularis mucosa of jejunum of patient 2 is filled with lipid deposits of various sizes. Contractile elements in this cell have been displaced to the periphery where they form a narrow zone. Eketron-lucent areas (arrowheads), presumed to represent lipid particles, are present in the extracellular spaces (x 13,500). 2930S~~~~~~~0

Figs 29 and 30-Parts of small nerves in jejunal submucosa of patient 2, showing extensive lipid deposits within the cytoplasm of Schwann cells. Some of the axons (30) contain small, lucent area (arrowheads) which may represent lipid deposits (X 16,200). a

-. 1% -s-~ s- v v -i

Fig 31-Low magnification view of foam cells in tonsil of patient 15. Many of the lipid deposits in these cells have angular shapes and are confluent with Figure 21 (x 6000). extensively Compare 1' i:

I :t

io lip. :, 'Ior. e. / a

AI... .. w '. Fig 32-Predominant type of foam cell in bone marrow of patient 15 contains large extensively coalescent lipid deposits with angular profiles. Note projections of cell surfaces; indentations caused by lipid deposits; and masses of debris. Compare with Figures 21 and 31 (x 9250). r t

-17-r. v

Fig 33-Foamy histocyte in bone marrow of patient 15 is adjacent to a neutrophil leuko- cyte and contains moderate numbers of lipid droplets and several fragments of phagocytized erythrocytes (x 11,500). 35 A

Fig 34Foam cell in bone marrow of patient 15 differs in structure from cells shown in Figures 32 and 33. This foam cell contains lipid deposits of various sizes and different degrees of electron density (X 8500). Fig 35High magnification view of part of cell shown in Figure 34. Electron-lucent, coalescent deposits are interspersed with amorphous or lamellar electron-dense material and with deposits which are moderately dense and contain small, clear spaces. All of these structures are free in the cytoplasm (X 34,500). t - * '11.

6.. -W N .. " A .0t, ,,

A :- 6 4 4, c. 1. -. Om 4 49 .6---

e DI 4)h

IV A *k-.

A * or ;,

Edg t*s Fig 36-Part of another foam cell in bone marrow of patient 15 showing large phagocytic vacuoles filled with lipid deposits of various sizes and with amorphous, electron-dense material. Many of the lipid deposits within the vacuoles are coalescent. Some small lipid particles (arrowheads) are free in the extracellular space (x 30,250). 38 iLw

Figs 37 and 38-Views of two foamy histiocytes in bone marrow of the father (clinically considered to be an obligatory heterozygote for Tangier disease) of patient 16. 37-Part of elongated cell which contains extensively coalescent deposits of lipid, fragments of phagocytized erythrocytes, and clusters of small lipid particles is shown. The structure of this cell resembles closely that of the cell (from patient 15) shown in Figure 32 (x 13,000). 38-Large lipid deposits and clusters of smaller lipid particles are surrounded by ferritin- like particles. These clusters are similar to those shown in Figure 36 (x 33,000). ; W: i -4 1: It 111. W. iat. I'..~~~~.. T411-* '-A. 'I VA

*. * W- J;@" r A i~0-f*-

-1 t

I.-

!t , r-I .11 .4, I 1. I -k

yl-; r'l ;, I .1 irj

Fig 39-Snall nerve in skin from patient 4 shows marked increase in endoneurial collagen; lipid deposits are present in Schwann cells surrounding unmyelinated and myelinated axons. A few axons contain accumulations of glycogen granules; others have small lipid droplets. Note marked thickening of basement membranes of Schwann cells and perineurial cells (x 5000). Fig 40Small cutaneous nerve from patient 5 shows changes, including fibrosis, lipid deposition in Schwann cells and thickening of basement membranes, which are similar to those in patient 4. Compare with Figure 39 (X 8000).