Macrophages, macrophage foam cells, and eccentric intimal thickening in the coronary arteries of young children

Arherosclerosis, 64 (1987) 91-108 Elsevier Scientific Publishers Ireland.

91 Ltd.

ATH 03904

Macrophages, macrophage foam cells, and eccentric intimal thickkning in the coronary arteries of young children H.C. Stary Louisiana State Unioersity, School of Medicine in New Orleans, New Orleans, LA (U.S.A.) (Received 13 May, 1986) (Revised, received 4 September, 1986) (Accepted 5 September, 1986)

summary We surveyed the incidence and location of macrophages and macrophage foam cells in the coronary artery intima of 63 children that died in the first 5 years of life. We related the data on macrophages and macrophage foam cells to intimal smooth muscle cells and to measurements of intima : media area and thickness. All morphometric data were obtained from coronary arteries that were fixed by perfusion with glutaraldehyde under pressure, embedded in Maraglas, and cut into l-pm cross-sections, and 65nm fine sections. Coronary artery intima was always thicker (eccentric thickening) at bifurcations in the half of the circumference opposite to the flow divider. This was true for both male and female children. The remaining part of the coronary artery intima was less thick (diffuse thickening). Both types of intimal thickening were composed of an inner layer in which glycosaminoglycan ground substance predominated and a deeper musculoelastic layer. Fifty-nine children (94%) had intimal macrophages. Twenty children also had macrophage foam cells. Of 33 children aged to 8 months, 15 (45%) had macrophage foam cells. Of the 30 children older than 8 months, 5 (17%) had macrophage foam cells. Macrophages and macrophage foam cells occurred in the GAG-rich layer of the intima as isolated cells. In 5 infants macrophage foam cells occurred also as clusters of many cells. Macrophages were more numerous in cases that also had macrophage foam cells. Macrophages were 6 times, and macrophage foam cells 5 times more numerous in eccentric intimal thickening than in diffuse intimal thickening.

Key

words: Atherosclerosis; phages;

Smooth

Coronary artery; muscle cells

Eccentric

intimal

thickening;

Electronmicroscopy;

Macro-

Introduction This work was supported by the National Institutes of Health, Grant HL-22739. _ Correspondence address: Herbert C. Stary. M.D.. L.S.U. School of Medicine, 1901 Perdido Street, New Orleans, LA 70112, U.S.A.

OO21-9150/87/$03.50

(0 1987 Elsevier Scientific

Publishers

Ireland,

Past studies of the coronary arteries of young children reported the existence of intimal thickening. The cells that composed the intima were not Ltd

92 specifically investigated. The presence of either macrophages, histiocytes, or macrophage foam cells was mentioned in only three studies of infant coronary intima [l-3]. In the intimal fatty streak and atheroma of adolescents and adults, macrophages and macrophage foam cells were more often reported [4-lo], and the literature on macrophages in experimental atherosclerosis is prolific. In 1913, the Journal of the German Pathological Society published the now classic paper by Anitschkow [ll] in which macrophages and macrophage foam cells were reported as a cellular component of the aortic intimal lesions of rabbits that had been eating food rich in cholesterol. Macrophages were then called phagocytes. In his drawings of the intimal lesions, Anitschkow pictured them (some in mitosis) and clearly differentiated them from modified (smooth) muscle cells (it was he who first used the term modified for lesion muscle cells). Subsequently, many investigators recognized macrophages and macrophage foam cells as components of the experimental intimal lesions of rabbits [12,13], rats [14-161, pigeons [17], pigs [18,19], and nonhuman primates [20-251. Cells of intimal lesions classified as macrophages and macrophage foam cells on the basis of their morphologic features have histochemical [26,27] and immunological [28-331 markers that are characteristically associated with macrophages. While there is general acceptance that macrophages participate in the evolution of atherosclerotic lesions, little more than that is known. Quantitations do not exist and basic questions that remain unanswered concern the relative contributions that macrophages make to the various types of intimal lesions and the part they play in lesion evolution and progression. The question of the presence and frequency of macrophages in normal (nonatherosclerotic) arterial intima, and their relation to eccentric thickening had either not been asked or not pursued. In this paper we report the incidence of macrophages and macrophage foam cells in human coronary artery intima in the first 5 years of life, and we relate this to the number of smooth muscle cells. We map the distribution of eccentric intimal thickening in the proximal coronary artery, delimit it from diffuse intimal thickening, and relate

macrophage and macrophage foam cell frequency to the presence and degree of intimal thickening. To put the present data in small children into perspective, we include some preliminary data on macrophage foam cells in coronary artery intimal lesions of the population aged from 5 to 29 years. Methods

The work reported here is part of a continuing study of the coronary arteries and aortas of human subjects that die between fullterm birth and 29 years. From 1979 to 1986 we obtained 1140 cases within this age range. We fixed the coronary arteries of 560 cases by perfusing them under pressure because the relatively short interval between death and autopsy of these cases (generally 10 hours or less) led us to expect tissue preservation adequate for study by high-resolution light microscopy and electron microscopy. Of the 560 cases with pressure perfusion-fixed coronary arteries, 70 were within the first 5 years of life. The 63 children reported here are those among the 70 cases in which all coronary artery sections were technically satisfactory so that all measurements and cell counts needed for this study could be made. Age and sex are shown in Fig. 3 (see p. 95). Causes of death are summarized in Table 1. Serum lipid levels were not determined in these children while they were alive. We obtained blood for serum cholesterol determination postmortem from 22 of the 63 children. High or low values did not correlate with the presence or absence of macrophage foam cells in the coronary artery intima. We prepared the coronary arteries by the following method: after the heart with the ascending aorta had been removed from the body, we connected the ascending aorta to a plastic tube which, at the other end, was connected to a container with 3% phosphate-buffered glutaraldehyde, elevated 135 cm above the level of the heart. The fixative was allowed to run into the coronary artery system through the ascending aorta by force of gravity. About 500 ml of glutaraldehyde were needed to obtain perfusion for about 20 min. To prevent loss of the fixative through the openings in the atria, we closed these, but to permit coronary flow and drainage, we kept the cut end of the

93 TABLE CAUSES

1 OF DEATH

BY AGE AND

Causes of death

BY PRESENCE Number subjects

OR ABSENCE

of

OF INTIMAL

MACROPHAGE By intimal

BY age (yr)

Present

Absent

12 4 5 2 5 4 3 43

19 1 2 2 1

0 12 5 5 1

3 3 2

3 2 2

11 1 3 2 0 1 1 1

Total number

63

33

30

20

flow

divider

(M0FC)

M0FC

l-4

19 13 7 7 2 6 5 4

pulmonary artery open. Ideally, this system produces an intracoronary pressure of 100 mm Hg. After perfusion we removed the unopened left

CELLS


Sudden infant death syndrome Accidents of various kinds Homicides Congenital malformations Inborn errors of metabolism Aspiration into lungs Pulmonary infections Infections other than lungs of subjects

FOAM

8

coronary artery from the heart. We used a dissecting microscope to clean the outside of fat, to sketch the gross layout and branching pattern, to measure the bifurcation angles, and to cut the proximal part into 5 consecutive tubelike portions, each measuring from 3 to 5 mm in length (Fig. 1). Length depended on age (that is, on heart size). We confined our studies of the perfusion-fixed coronary arteries to this 15-25 mm segment be-

wall

L.A

Fig. 1. The drawing indicates how we divided the proximal part of the left coronary artery into 5 separate portions before embedding each portion in Maraglas. Because of anatomical variation, some portions were not available in some subjects. On the other hand, we sometimes obtained additional portions not indicated in this drawing. Thus, the proximal part of the circumflex and intermediate branches, and more distal segments of the left anterior descending branch were obtained in some subjects. Each portion was cut into multiple l-pm crosssections. To obtain, in the sections, identification marks of the flow divider and anterior walls, and also to earmark the proximal end of each portion, we incised each portion as shown before embedding it in Maraglas.

Fig. 2. Sketch of a cross-section of the normal human coronary artery at the L.A.D. 1 level. The distribution of the thickness of the intima around the circumference of the coronary artery is typical for this location. Eccentric thickening of the intima (E.T.) is typically located opposite the flow divider wall, while diffuse thickening (D.T.) involves the remainder of the circumference. We determined the position of E.T. and D.T. from the marks which we had incised into the flow divider and anterior walls of each portion of the artery before we embedded it in plastic. The relationship of the two layers of the intima (GAG-rich and musculoelastic) to each other and to the media are also shown.

94

cause atherosclerotic lesions, if they occur at all, occur in this location with predictable regularity [34]; it is here also that lesions become most often occlusive [35]. This precisely defined coronary segment is therefore ideally suited for the close study of the spectrum of lesions from their earliest inception through intermediate stages of evolution to stages of great complexity and clinical significance. We marked the anterior wall and the flow divider wall of each tubelike portion with small incisions (Figs. 1 and 2). In subsequent histological cross-sections, these cuts indicated the location of eccentric intima thickening and the location of lesions in relation to the flow divider wall and anterior and posterior walls of the artery. The portions were then immersed in 3% buffered glutaraldehyde (for a total fixation time of 2 h), washed in buffer, post-fixed in osmium, dehydrated, and embedded upright in Beem capsules containing Maraglas. From each tubelike portion we cut multiple l-pm cross-sections with glass knives. In many cases we made semiserial crosssections through the entire block. We stained sections with toluidine blue, basic fuchsin, and sodium borate. We obtained morphometric data from l-pm cross-sections by projecting them on a horizontal digitizer board and by digitizing the areas of the media, the glycosaminoglycan (GAG)-rich intima layer, the musculoelastic intima layer, and total intima. In addition to areas, we measured, at equidistant points around the circumference, the thickness of these layers. The border between the GAG-rich and musculoelastic layers of the intima was sometimes difficult to digitize because it was sometimes ill defined. The GAG-rich intima extended into the adjacent musculoelastic intima, while peninsular projections of densely packed smooth muscle cells extended into the GAG layer from the musculoelastic intima. In our measurements and cell counts we considered projections of GAG into the musculoelastic layer as GAG intima. Similarly, it was sometimes difficult to digitize the line of demarcation between intima and media since the internal elastic lamina was often incomplete or absent. This was more often the case in segments with eccentric thickening than in segments with diffuse thickening. To digitize such cross-sections we reconstructed a com-

plete internal elastic lamina by connecting incomplete segments or, in their absence, by choosing the most likely demarcation line. Data were stored in a ProFile Extension hard disk and analyzed with AppleWorks and own programs by an Apple IIe computer. We identified intimal macrophages, macrophage foam cells, and smooth muscle cells in the l-pm cross-sections by light microscopy with Nikon NCG 60 X and 100 X (dry) objectives. To count the cells we used an eyepiece grid. Although accurate identification of cells may require electron microscopy, the exclusive use of electron microscopy would not have allowed us to study cells around the entire coronary circumference in each case or to count the number of cells appropriate to our study. We used electron microscopy selectively to check identification and quantitations of macrophages, macrophage foam cells and smooth muscle cells in limited regions of the intima. For electron microscopy we chose the parts of the coronary cross-sections with the largest number of macrophages and macrophage foam cells as determined by microscopy of the 1-,um cross-sections. The 65-nm sections were obtained with a diamond knife, stained with uranyl acetate and lead citrate, and photographed in a Philips 201 or 300 electron microscope. Some intimal cells we could not classify by light or by electron microscopy. We believe, however, that we correctly indentified the majority. Results Normal intima: thickening

eccentric

thickening

and

diffuse

Every coronary artery, from the first week of life on, contained two basic intima patterns to which we applied the terms eccentric thickening and diffuse thickening. The patterns were continuous and complemented each other. Eccentric intima thickening (eccentric thickening) was closely localized to bifurcations (Figs. 2 and 5-8), and generally limited to the one half of the coronary circumference that was roughly opposite the flow divider wall. It appeared as a crescent-shaped increase in intima thickness which at its midpoint could attain several times the thickness of the media. It is also known as cushion

95 [1,36], pad [37], or bolster [38], but it resembles all of these only in a artery that has been fixed after it has collapsed. McMillan, who discovered traces of lipid in some, described them as mucoid fibromuscular plaques [39], and Velican and Velican [40] labeled some larger eccentric thickenings of 6-lo-year-old children as lipid-free fibromuscular plaques. I have avoided the word plaque for these structures because the term has the connotation of a pathological process. In coronary arteries,

/:$I

Infant

Infant

with

M$FC

without

M$FC

??

1.5 -

. 0

0

0 0

l.O-

the most prominent eccentric thickening was that associated with the left main bifurcation. It reached from the main branch, just proximal to the bifurcation, into the left anterior descending branch. Maximal thickening was usually present at the origin of the anterior descending, growing gradually smaller and tapering into diffuse thickening a variable distance distal to the bifurcation. At the next downstream bifurcation the above pattern was repeated. Diffuse intimal thickening (diffuse thickening) was more extensive and occurred in regions that were not involved with eccentric thickening (Figs. 2 and 7). Here, intima thickness was moderate and relatively uniform, less than or just equaling that of the media. At and near a bifurcation, diffuse thickening was continuous with eccentric thickening, involving the one half of the coronary artery circumference not involved with eccentric thickening. Away from bifurcations it encompassed the entire circumference. In the literature the term diffuse intimal thickening is often used. It is probably equivalent to the terms fibromuscular and musculoelastic thickening [34,41]. Some authors

0.5

to-

Year5

Months Age

T

.o_

P E

Fig. 3. In this graph, the 63 infants are arranged by age and by the thickness of their intima. Intima thickness is expressed as the ratio of the area of the intima to the area of the media. A value of 1.0 indicates that the areas of intima and media were equal. A value higher than 1.0 indicates that intima exceeded media. Measurements of area shown here represent the mean of measurements obtained from cross-sections at the main, bifurcation and L.A.D. 1 levels (see Fig. 1). In some subjects a main portion was absent. In these cases the measurements are those obtained at the bifurcation and at the L.A.D. 1 level only. The great variation in intima thickness between subjects is chiefly attributable to individual variation in the degree of eccentric thickening. Diffuse thickening and media thickness varied less between individuals, Our measurements (not included here) convey the uneveness in intima thickness around the circumference of each individual coronary artery. Each infant is coded to show sex and whether or not macrophage foam cells (MBFC) were present. The data indicate that presence or absence of macrophage foam cells in the intima is not related to differences between infants in intima area (actually area of eccentric intima thickening) or to a difference in sex.

: 2 0.5 .-; +c P 2

Cases without M$FC

Fig. 4. The mean intimal thickness) phage foam cells coronary arteries macrophage foam macrophage foam

Cases with M@FC

(*SE) intima to media ratio (our index of by sex and presence or absence of macro(MBFC). The means were obtained from of 25 female, and 18 male infants without cells; and 8 female, and 12 male infants with cells,

Fig. 5. Low-power light micrograph of the left coronary artery at the level of the main bifurcation. Two ridges (FD) in the ventral and dorsal walls will unite distally to form the apex of the wall that will separate blood flow and that will form the inner (flow divider) walls of the L.A.D. and circumflex daughter vessels. The emerging L.A.D. branch contains eccentric thickening (ET) opposite the evolving flow divider (FD). From a 6-month-old boy who was murdered. Case No. 949 (P-1949); l-pm section, X 64. Fig. 6. Cross-section of the main trunk of the left coronary artery. Eccentric thickening begins proximal to a bifurcation. It is larger and often limited (as in this case) to the side of the artery wall that will, more distally, continue as the outer wall of the L.A.D. branch. From a 3-month-old male infant. The cause of death was sudden infant death syndrome. Case No. 722 (P-1722); l-pm section, X 40. Fig. 7. The outer wall at the L.A.D. 1 level shows the eccentric thickening (ET) which, to a variable degree, was present in this location in all subjects and at all ages. Diffuse thickening (DT). From a 2-year-old boy who died of aspiration into the lungs. Case No. 585 (P-1585); l-pm section, x40.

Fig. 8. Eccentric thickening at the L.A.D. 1 level, magnified to show the inner GAG-rich layer (gag), the underlying musculoelastic layer (me), and just below the endothelial cell layer, macrophages that are not foam cells (arrows). M = media: A = adventitia. From a l&month-old boy who was murdered. Case no. 476 (P-1476); l-pm section, x 170. Fig. 9. Detail of Fig. 8. Macrophages without visfble lipid droplets intima layer (me): e = endothelial cells. l-pm section, x 880. Fig. 10. Detail of Fig. 8. Isolated macrophage Musculoelastic intima layer (me); e = endothelial

(arrows)

occupy

without visible lipid droplets cell. l-pm section, x 880.

the GAG-rich

(arrow)

layer of the intima.

in the GAG-rich

layer

Musculoelastic

of the intima.

98

Fig. 11. A macrophage that is not a foam cell is just below the endothelial surface (E) within the GAG-rich intima thickening at the L.A.D. 2 level. The cell contains isolated lipid droplet inclusions and lysosomes. Electron micrograph, x 17800.

matrix of the eccentric Same case as in Fig. 8.

Fig. 12. A macrophage that is not a foam cell within the GAG-rich matrix of the eccentric intima thickening at the L.A.D. 2 level. The cell contains small, isolated lipid droplet inclusions and lysosomes. Same case as Fig. 8. Electron micrograph, x 13 200.

99 TABLE

2

NUMBER OF INTIMAL BUT WITHOUT M0FC

MB IN 39 INFANTS

WITH

M0

Means and ranges are those of all subjects included in each l-year age group. For each subject and location, the cell number was obtained by counting intimal MB around the entire coronary artery circumference of one l-pm thick crosssection. Age

Number

Number

(yr)

of subjects

Main

Bifurcation

L.A.D. 1 L.A.D. 2 L.A.D.

cl

18

4 o-12

5 o-17

5 o-15

1

2

3

4

5

2

6

6

4

MB (mean and range)

3 O-10

1

2

1

7

o-

9

o-

2

0-

8

o-2

4 2-

7

4 l-

9

2 o-

4

1 o-

3

2 o-5

o-

5

o-

7

o-

4

o-

2 o-

3

2 o-

9

1 o-

3

1 o-4

3

1

3

3 O-6

o-

2

4

of intimal

1

4 4

l-8 1 1

may not distinguish between the diffuse and the eccentric patterns, applying anyone of the terms I listed here and further above to both patterns. The area measurements used in the graph (Fig. 3) comprise the areas of both eccentric and diffuse thickening, which together make up the circumference. Differences between the subjects in TABLE

intima area could, therefore, be caused by variation in either eccentric or diffuse thickening. But since we also measured intima thickness at several points of the circumference, we know that the differences were caused by changes in the degree of eccentric thickening. Similarly, the decrease in intima area which occurred distal to a bifurcation reflected the gradual loss of eccentric thickening rather than a change in diffuse thickening. There was no difference between male and female infants in the ratio of intimal area to medial area (Fig. 4). Both eccentric and diffuse thickening were composed of two main layers which differed in density and nature of smooth muscle cells, and in density and nature of the extracellular matrix. The inner (luminal) layer was rich in glycosaminoglycan (GAG) matrix and poor in elastic fibers. GAG sometimes formed pools, often extensive in the youngest infants (see Fig. 13). I, therefore, call it the GAG-rich layer. Here, smooth muscle cells were loosely arranged and more often of the RER-rich type. The thicker, underlying (musculoelastic) layer was poor in GAG and rich in elastic fibers. Smooth muscle cells were densely packed, in an orderly arrangement, and of the myofilament-rich type. In these young children, neither the smooth muscle cells in the GAG-rich intima layer, or those in the musculoelastic intima layer, nor those in the media contained cyto-

3

NUMBER INFANTS

OF M0, MIZIFC, AND THAT HAD INTIMAL

SMOOTH MUSCLE CELLS (SMC) IN THE GAG-RICH LAYER OF THE INTIMA OF 15 MOFC AND THAT DIED IN THEIR FIRST YEAR OF LIFE ( < 1 YEAR AGE GROUP)

Means and ranges are of all cases combined in this age group. In each subject and in each of the 5 coronary locations we counted M0, MBFC, and SMC in the GAG-rich layer of the intima around the entire coronary circumference in one, l-pm thick, cross-section. Five of the 20 children with intimal MBFC were between the age of 1 and 4 years. Data from these 5 older children are not included in this table. Main MO

MBFC

SMC

mean number range

10 4-23

Bifurcation 12 1-31

L.A.D. 1

L.A.D. 7 1-14

8 2-17 3 O-10

mean number

5

7

4

o-19

O-40

o-12

3 o-14

Proportion of cells that are MB or MBFC

L.A.D.

8 2-16

range mean number range

2

277 117-610

419 176-752

324 126-681

261 89-547

252 129-387

6.2%

5.0%

3.9%

5.0%

6.7%

3

Fig. 13. Overview of eccentric intimal thickening at the L.A.D. 1 level. The GAG-rich layer (gag) of the intima separated from the musculoelastic layer of the intima (me). Such overlap of layers is characteristic of eccentric intimal young infants. Single macrophage foam cells (small arrows) and a cluster of many macrophage foam cells (large arrows) accumulations of GAG. The remaining cells in this picture are intimal smooth muscle cells. The internal elastic lamina which underlie the intima are not included in this picture. lu = lumen. From a l-month-old male infant who died in an No. 135 (P-1135); l-pm section, x420. Fig. 14. Detail of Fig. 13. Three isolated Fig. 15. Detail of Fig. 13. A cluster

macrophage

of macrophage

foam cells (arrows) foam cells (arrows)

occupy occupies

a pool of GAG a pool of GAG

matrix. matrix.

X 880. x 880.

is nc ,t clearly thickening in occupy larger and the media accident. Case

101

Fig. 16. Detail of Fig. 13. Electron micrograph of macrophage of the cluster pictured in Figs. 13 and 15. X51ot).

foam cells (F) occupying

the GAG matrix (gag). The four cells are part

Fig. 17. Detail of Fig. 16. The fine, cobweb-like structure of the ground substance is characteristic of the morphological of glycosaminoglycans. Portions of two macrophage foam cells (F). Eelctron micrograph. x 34500.

appearance

102

plasmic accumulations of lipid droplets. Eccentric thickening differed from diffuse thickening only by a greater thickness of each layer and by a larger number of macrophages.

Frequency and location of macrophages

The intima of 59 (94%) of the 63 c:hil contained macrophages that were not foa .m cells

Fig. 18. A detailed view of one of the macrophage foam cells shown in Fig. 16. A portion of an adjacent macrophage foam ccl 1 is in the lower left comer. N - nucleus; v = plasma membrane microvilli; L = lipid droplet inclusions; gag = glycosaminoglycan ma1trix of the intima. Electron micrograph, x 12600.

103

within the limits of the coronary artery segment we studied. Macrophages were limited to the part of the GAG-rich layer of the intima that was immediately subendothelial (Figs. 8-12). While macrophages occurred around the entire coronary artery circumference, their density was greater in eccentric than in diffuse thickening. Even so, macrophages were relative thinly dispersed as isolated cells, and not arranged as dense clusters of cells as was sometimes the case with macrophage foam cells. The number of macrophages in the coronary intima varied from one case to the next. Macrophages ranged from 0 to 17 per complete l-pm cross-section in 39 cases that were without foam cells (Table 2). Macrophages were twice as numerous in cases that had, in addition, macrophage foam cells (Table 3). Tables 2 and 3 give counts made around the entire coronary circumference, not distinguishing between macrophage frequency in eccentric and in diffuse intima thickening. When we recounted macrophages to determine their frequency in eccentric versus diffuse thickening we found that intima in the half of the coronary circumference that included eccentric thickening contained 6 times more macrophages than the opposite half of the circumference which included diffuse thickening. Frequency and location of macrophage foam cells

The intimas of 20 of the 63 infants had macrophage foam cells within the limits of the coronary artery segment we studied. Of 33 infants aged between fullterm birth and 8 months, 15 (45%) had macrophage foam cells in coronary artery intima. Subsequently, in older infants, macrophage foam cell incidence declined. Of the 30 children older than 8 months, only 5 (17%) had macrophage foam cells in coronary artery intima. Macrophage foam cells occupied the GAG-rich layer of the intima (Figs. 13-18). They occurred around the entire coronary artery circumference, that is in both eccentric and in diffuse thickening, but they were 5 times as numerous in eccentric thickening. Usually macrophage foam cells were distributed as widely spaced isolated cells, but 5 infants had, in addition, macrophage foam cells clustered as groups of cells (Figs. 13, 15, 16). The

clusters always occurred in eccentric thickening. Cases with the higher concentration of macrophage foam cells were within the first 5 months of life. The youngest infant (8 days old) in our sample had clusters of macrophage foam cells. The pooled GAGS of eccentric thickening constituted the preferred spot for the accumulation of macrophage foam cells (Fig. 13). Figure 3 lists children according to their age and sex, and relates the presence or absence of macrophage foam cells in each child to the area of the intima in each child. Differences in area between children reflect differences.in the degree of eccentric thickening. The degree of eccentric thickening did not determine whether or not macrophage foam cells were present in the coronary arteries of an individual subject. A slightly greater intima area in cases with foam cells is suggested when the mean of all cases with foam cells is compared with the mean of all cases without (see Fig. 4). A small difference in area could be caused by the presence of the foam cells themselves. The number of macrophage foam cells ranged from 0 to 40 per complete l-pm coronary crosssection. In the GAG-rich layer there were from 1.2 to 1.8 macrophage foam cells per 100 smooth muscle cells. Macrophages and macrophage foam cells combined constituted from 3.9 to 6.7 cells per 100 smooth muscle cells (Table 3). The flow divider wall showed diffuse intimal thickening except for the apex. The thickness of the intima at the apex (see Fig. 5) approached the thickness of the eccentric thickenings on the opposing walls of the two daughter vessels. Nevertheless, neither macrophages nor macrophage foam cells were nearly as numerous in the flow divider apex as they were in eccentric thickening. Discussion Cause of macrophage foam cells

Macrophage foam cells were more often present and more numerous in the intima of infants that died soon after birth. This observation suggests that macrophage foam cells in the intima of infants and their subsequent regression reflect short-term changes in infant nutrition and, thus, changes in plasma lipids very early in life. Analysis of serum lipids in blood taken from the umbili-

104 cal cord at birth and in blood obtained 4 days and 6 weeks later [42] or 3 months later [43] or three years later [44] showed a dramatic increase within days following birth. Although the final levels were below those that are associated with the development of deposits of lipid in the intima in adults, the sudden rapid increase in serum cholesterol after birth rather than the absolute level might have generated the foam cells we observed. Breast-fed infants generally have higher serum cholesterol levels than infants that receive a formula diet [43]. Perhaps this dietary difference accounted for the presence of foam cells in a portion of the children in our study. Postmortem serum cholesterol measurements obtained in a small number of infants in the present study did not correlate with presence or absence of foam cells. There is also some evidence suggesting that foam cells might be generated in intrauterine life, reflecting the plasma lipid levels of the mother. Pregnant women have higher serum cholesterol levels [45], and transfer of cholesterol across the placenta is possible, at least in experimental primates [46]. Unfortunately, the serum cholesterol levels of the mothers of our children are not known. Foam cells have been documented in the intima of pulmonary artery branches in infants that were born prematurely and had received lipid infusions (Intralipid@) [47]. Lipid infusions are used as a source of calories and of essential fatty acids in infants who cannot be adequately nourished by mouth. None of the infants in our study were premature, and none had received lipid infusions. Origin of intimal macrophages

Some deductions on the origin of intimal foam cells are possible by relating our data in young children to experimental studies. After the first 8 months of life, few small children had foam cells in the coronary intima, but macrophages that were not foam cells were inevitably present. It is likely, therefore, that intimal macrophages represent a resident population, available to internalize excess lipoprotein from their environment, and capable of at least some cell division when confronted with this situation. Evidence that macrophage foam cells can increase by proliferating locally, in the

intima, is strong. We reported mitotic foam cells with the ultrastructural characteristics of macrophages in the experimental intimal lesions of monkeys [23]. Recently, in these studies of young people, we also identified mitotic foam cells with fine structural features of macrophages in the coronary artery fatty streak of a 25-year-old man. Supportive evidence comes from a radioautographic and electron-microscopic study of human fibroatheromatous plaques [lo]. Here, Villas&i and Spagnoli found that cells with the features of monocytes and macrophage foam cells were labeled with tritiated thymidine. Furthermore, rereview of material from our own light-microscopic radioautography studies of experimental atherosclerotic lesions [48] indicates that the majority of the cells that were labeled with tritiated thymidine, were the round foam cells that, as new, electron-microscopic and immunocytochemical evidence confirms, are more often macrophages with lipid droplets than smooth muscle cells with lipid droplets. The question of whether monocytes and macrophages can divide had been controversial [49]. However, there is much evidence now that macrophages derived from circulating monocytes and resident macrophages divide in various tissues, in a number of situations. In inflammatory exudates, macrophages derived from circulating monocytes continue to divide [50-521, and proliferation of resident macrophages has been demonstrated in the lungs [49,53], in the liver [54], and in the peritoneal cavity [55-571. To what extent a resident intimal macrophage population is replenished and supplemented by monocytes that migrate into the intima from the circulation is unclear at present. Scanning and transmission electron-microscopic data suggest that monocytes can transmigrate the endothelium in significant numbers to populate the intima [17,18,25,58]. Some investigators believe that atherosclerotic lesions generate chemoattractants that recruit circulating monocytes to the intima [59]. Regression of macrophage foam cells

The decline, after the first 8 months, in cases with foam cells, and in the number of foam cells, suggests that foam cells that had been present earlier were removed, and that resident intimal

105

macrophages did not form foam cells much for several years thereafter. We have no evidence that foam cells were removed from the intima by transmigrating the endothelium. Since we saw some dead foam cells in the intima by electron microscopy, we assume that macrophage foam cells disappear because they die. Remnants of the dead cells did not accumulate in the intima of these young children; clearly, the rate of accumulation did not exceed the capacity of the arterial wall to clean itself. In rhesus monkeys in which food rich in cholesterol had caused massive foam cell accumulations, foam cells disappeared (we have electronmicroscopic evidence that they died in the intima) after dietary and serum cholesterol was drastically reduced. The interval between diet change and disappearance from the intima of most foam cells was 4-6 months [60]. This is an indication of the approximate life span of macrophage foam cells in monkeys. If the intimal macrophage foam cells of young children have a similar life span, then it is likely that the stimulus that caused them ceased or diminished very early in the first year of life. The statement that lipid can regress from the aortic intima of small children recurs throughout the literature. Although evidently true, past evidence has been anecdotal. Zinserling [61] who studied fatty streaks that were grossly visible in aortas stained with Sudan, was uncertain about whether they could regress. Nevertheless, Aschoff [62] in referring to Zinserling’s work, wrote that by the time of puberty, lipid earlier deposited in the mitral valve had more or less disappeared, while at the root of the aorta fat deposits might still be recognizable, thus, implying regression of what he termed ‘suckling infants atherosis’. Aschoff did not claim regression of ‘pubertal atherosis’, that is, of fatty streaks in the aorta of older children. Some evidence of the reduction in size, or of the disappearance, of human atherosclerotic lesions has been compiled in several recent literature reviews [63,64]. Solid evidence comes only from animal experiments [60,65]. Participation of macrophages and macrophage foam cells in the development offatty streaks and atheroma Data on the development of fatty streaks and atheroma may serve to put the intimal macro-

phages of young children into perspective. These data too come from the 560 cases with pressureperfusion fixed coronary arteries available so far in our studies of the first 29 years of life [66]. Detailed data of fatty streaks and atheroma will appear in separate publications. The study of the 63 youngest children revealed a preference of macrophages and macrophage foam cells for eccentric thickening. This special relationship to eccentric thickening continued when macrophage foam cells re-emerged in greater number in older children as a component of fatty streaks. The term fatty streak describes a lesion in which intimal lipid becomes visible grossly. Microscopically, lipid droplets overload the cytoplasm not only of macrophages but also of smooth muscle cells, and small lipid particles are scattered in the extracellular space of the intima. In the proximal coronary artery, lipid deposits having the microscopic configuration of fatty streaks were heaviest in eccentric thickening. Nevertheless, they were, in this location, inconspicuous from the endothelial surface. This was so because lipid droplets accumulated in smooth muscle cells that were located in the deep part of the GAG-rich intima layer and macrophage foam cells and extracellular lipid occupied the same region. The close relationship between macrophage foam cells, eccentric thickening, and lipid accumulation continued in the subsequent developmental stage of atherosclerosis, the atheroma. The term atheroma describes an atherosclerotic lesion that contains a massive, confluent core of extracellular lipid. In the young population, lipid cores (i.e. atheroma) occurred only in eccentric thickening. Here they developed in the deep layer of eccentric thickening just below that in which macrophage foam cells predominated. Because there were dead cells within the foam cell layer, we conclude that lipid cores formed through the accumulation of remnants of many generations of macrophage foam cells. As a fatty streak within an eccentric thickening changed into atheroma and as atheroma enlarged, the proportion of the existing (and possibly also of newly proliferated) smooth muscle cells that acquired lipid droplet inclusions increased in relation to the number of macrophage foam cells. Thus, smooth muscle cells became the numerically

106 dominant cell type that was laden with lipid droplets. Still, macrophages appear to be the first line of defense in the intima, being the first to accumulate, store, and digest lipid droplets. Because of their apparent capacity for rapid internalization and degradation, and because of their rapid turnover (abundant immigration, replication in the intima, short life-span), macrophages certainly account for the uptake of most of the lipoprotein, while at the same time their remnants, when produced rapidly over extended periods, constitute the bulk of the cores of atheroma. Fatty streaks emerged in the coronary arteries in the 5-year period between 10 and 14 years. We found fatty streaks in half of all children in that age group [66]. This proportion is similar to that of infants with intimal foam cells in the first 8 months of life. In spite of similarities in incidence, morphology, and location, we do not regard the intimal macrophage foam cells of infants as a direct precursor of fatty streaks. The absence of continuity between the foam cells of infants .and the fatty streaks of older children (as seen in the foam cell-poor interval after the first 8 months of life) speaks against it. We have no evidence that infants with intimal foam cells will be the children that develop fatty streaks and atheroma in the same location.

Our data support investigators who have studied flow and mechanical wall stresses and who consider eccentric thickening to be an anatomical structure that develops as a response to intrinsic hemodynamic situations [67,68]. In this view, eccentric thickening is a physiologic adaptation to certain hemodynamic forces to which parts of the vasculature are exposed from the very development of the vascular system. Genetic differences in size and layout of coronary arteries presumably cause individual variation in hemodynamic patterns, and thus determine thickness, extent, and the precise location of eccentric thickening. Perhaps individual differences of a constitutional or hormonal nature also play a role.

The significance of eccentric thickening We determined that isolated macrophage foam cells and fatty streaks favored eccentric thickening, and that atheroma was limited to it, independent of the degree of thickness or of the precise extent of eccentric thickening. We infer from this that forces related to the existence of eccentric thickening also cause a greater influx of plasma lipoprotein into these sites, and that macrophages become active when the concentration of certain lipoproteins is high in the intima. We found eccentric thickening in the coronary arteries of all subjects. Clearly, since eccentric thickening is universal in man from birth, it is a normal, not a pathological structure. It is incorrect to think of it as an atherosclerotic or even arteriosclerotic lesion since it occurs in man whether or not arterial disease is superimposed on it later, and since by itself it never is obstructive to flow.

Fangman, R.J. and Hellwig, C.A., Histology of coronary arteries in newborn infants, Abstract, Amer. J. Path., 23 (1947) 901. Moon, H.D., Coronary arteries in fetuses, infants, and juveniles, Circulation, 16 (1957) 263. Jaffe D., Hartroft, W.S., Manning, M. and Eleta, G., Coronary arteries in newborn children, Acta Paediat. Stand., Suppl. 219 (1971) 1. Duff, G.L. and McMillan, G.C., Pathology of atherosclerosis, Amer. J. Med., 11 (1951) 92. Still, W.J.S. and Marriott, P.R., Comparative morphology of the early atherosclerotic lesion in man and cholesterolatherosclerosis in the rabbit. An electron-microscopic study, J. Atheroscler. Res., 4 (1964) 373. Geer, J.C., Fine structure of human aortic intimal thickening and fatty streaks, Lab. Invest., 14 (1965) 1764. Adams, C.W.M. and Bayliss, O.B., Detection of macrophages in atherosclerotic lesions with cytochrome oxidase, Brit. J. Exp. Path., 57 (1976) 30. Stary, H.C., Macrophages in coronary artery and aortic intima and in atherosclerotic lesions of children and young adults up to age 29. In: F.G. Schettler et al. (Eds.), Atherosclerosis VI, Springer-Verlag, New York, Heidelberg, Berlin, 1983, pp. 462-466.

Acknowledgments

My thanks are due to those in our laboratory who made this work possible: James Stoll quantitated intimal cells in coronary cross-sections; Rhett Hubley and Fayne St. John digitized areas and thickness of coronary cross-sections; Elizabeth Donnell and John Pleshinger cut and stained the l-pm cross-sections and the fine sections; Patricia Mangiaracina prepared the manuscript. References

107 9 Goldstein, J.L. and Brown, MS., Familial hypercholesterolemia. In: J.B. Stanbury (Ed.), The Metabolic Basis of Inherited Disease, 5th edition, McGraw Hill Book Company, New York, 1983, pp. 672-730. 10 Villas&i, S. and Spagnoli, L.G., Autoradiographic and ultrastructural studies on the human fibro-atheromatous plaque, Atherosclerosis, 48 (1983) 95. 11 Anitschkow, N., Uber die Ver&denmgen der Kaninchenaorta bei experimenteller cholesterinsteatose, Beitr. Path. Anat. Allgem. Path., 56 (1913) 379. 12 Duff, G.L., McMillan, G.C. and Ritchie, A.C., The morphology of early atherosclerotic lesions of the aorta demonstrated by the surface technique in rabbits fed cholesterol. Amer. J. Path., 33 (1957) 845. 13 Cavallero. C., DiTondo, U., Mingazzini, P.L., Pesando. P.C. and Spagnoli, L.G., Cell proliferation in the atherosclerotic lesions of cholesterol-fed rabbits, Atherosclerosis, 17 (1973) 49. 14 Still, W.J.S. and O’Neal, R.M., Electron microscopic study of experimental atherosclerosis in the rat, Amer. J. Path., 40 (1962) 21. 15 Gerrity, R.G. and Cliff, W.J., The aortic tunica media of the developing rat, Part 1 (Quantitative stereologic and biochemical analysis), Lab. Invest., 32 (1975) 585. 16 Joris. I., Zand, T.. Nunnari, J.J., Krolikowski. F.J. and Majno, Cr.. Studies on the pathogenesis of atherosclerosis, Part 1 (Adhesion and emigration of mononuclear cells in the aorta of hypercholesterolemic rats), Amer. J. Path., 113 (1983) 341. 17 Lewis, J.C., Taylor, R.G. and Jerome, W.G., Foam cell characteristics in coronary arteries and aortas of White Cameau pigeons with moderate hypercholesterolemia. Ann. N.Y. Acad. Sci., 454 (1985) 91. 1X Gerrity, R.G.. The role of the monocyte in atherogenesis, Part 2 (Migration of foam cells from atherosclerotic lesions), Amer. J. Path., 103 (1981) 191. 19 Daoud, AS., Jarmolych, J.. Augustyn. J.M. and Fritz, K.E., Sequential morphologic studies of regression of advanced atherosclerosis, Arch. Path. Lab. Med., 105 (1981) 233. 20 Geer, J.C.. Catsuhs, C., McGill Jr., H.C. and Strong, J.P.. Fine structure of the baboon aortic fatty streak, Amer. J. Path., 52 (1968) 265. 21 Stary, H.C.. Coronary artery fine structure in rhesus monkeys: The early atherosclerotic lesion and its progression, Primates in Med., 9 (1976) 359. 22 Stary, H.C., Strong, J.P. and Eggen, D.A., Differences in the degradation-rate of intracellular lipid droplets in the intimal smooth muscle cells and macrophages of regressing atherosclerotic lesions of primates. In: A.M. Gotto et al. (Eds.). Atherosclerosis V, Springer-Verlag. New York, 1980, pp. 753-756. 23 Stary, H.C. and Malinow, M.R.. Ultrastructure of experimental coronary artery atherosclerosis in cynomolgus macaques, Atherosclerosis, 43 (1982) 151. 24 Faggiotto, A., Ross, R. and Harker, L.. Studies of hypercholesterolemia in the nonhuman primate, Part 1 (Changes that lead to fatty streak formation), Arteriosclerosis, 4 !1984) 323.

25 Schwartz, C.J., Sprague, E.A., Kelley, J.L., Valente, A.J. and Suenram, C.A., Aortic intimal monocyte recruitment in the normo and hypercholesterolemic baboon (Pupio cynu cephulus), Virchow’s Arch. (Path. Anat.). 405 (1985) 175. 26 Fritz, K.E., Daoud, A.S. and Jarmolych, J.. Non-specific esterase activity during regression of swine aortic atherosclerosis, Artery, 7 (1980) 352. 27 Schaffner, T., Taylor, K., Bartucci, E.J., Fischer-Dzoga. K.. Beeson, J.H., Glagov. S. and Wissler. R.W.. Arterial foam cells with distinctive immunomorphologic and histochemical features of macrophages, Amer. J. Path., 100 (1980) 57. 28 Klurfeld, D.M., Interactions of immune function with lipids and atherosclerosis, CRC Critical Rev Toxicol., 11 (1983) 333. 29 Aqel, N.M.. Ball, R.Y.. Waldmann. H. and Mitchinson. M.J., Identification of macrophages and smooth muscle cells in human atherosclerosis using monoclonal antibodies. J. Path., 146 (1985) 197. 30 Vedeler. C.A., Nyland, H. and Matre, R., In situ characterization of the foam cells in early human atherosclerotic lesions, Acta Path. Microbial. Immunol. Stand. Sect. C., 92 (1984) 133. 31 Tsukada, T., Ross, R. and Gown, A.M.. Smooth muscle cell- and macrophage-specific monocolonal antibodies in the immunocytochemical analysis of human and rabbit atherosclerotic lesions, Arteriosclerosis, 5 (1985) 497a. 32 Watanabe, T., Hirata, M.. Yoshikawa. Y.. Nagafuchi. Y.. Toyoshima, H. and Watanabe, T.. Role of macrophages in atherosclerosis. Sequential observations of cholesterol-induced rabbit aortic lesion by the immunoperoxidase technique using monoclonal antimacrophage antibody, Lab. Invest., 53 (1985) 80. 33 Munro, J.M.. Van der Walt, J.D., Chalmers. J.A.C. and Cox, E.L., Immunopathological study of aortic fatty streaks, J. Path., 148 (1986) ill*. 34 Geer, J.C., McGill, Jr., H.C., Robertson, W.B. and Strong, J.P.. Histologic characteristics of coronary artery fatty streaks, Lab. Invest., 18 (1968) 565. 35 Berger. R.L. and Stary, H.C.. Anatomic assessment of operability by the saphenous vein bypass operation in coronary artery disease, N. Engl. J. Med., 285 (1971) 248. 36 Dock, W., The predilection of atherosclerosis for the coronary arteries, J. Amer. Med. Ass., 131 (1946) 875. 37 Stehbens, W.E., Focal intimal proliferation in the cerebral arteries, Amer. J. Path., 36 (1960) 289. 38 Minkowski, W.L.. The coronary arteries of infants. Amer. J. Med. Sci., 214 (1947) 623. 39 McMillan. G.C., The onset of plaque formation in arteriosclerosis, Acta Cardiol.. Suppl. XI (1965) 43. 40 Velican, C. and Velican D., The precursors of coronary atherosclerotic plaques in subjects up to 40 years old. Atherosclerosis, 37 (1980) 33. 41 McGill. Jr., H.C., Atherosclerosis: Problems in pathogenesia, Atheroscler. Rev., 2 (1977) 27. 42 Tan, M.H.. Dolphin, P.J.. Breckenridge. W.C. and Stewart. B., Changes in serum lipoproteins during the first six weeks of life, Clin. Invest. Med., 8 (1985) A169. 43 Gtirson. C.T. and Saner, G.. Serum cholesterol levels at

108 birth, in infancy and at the age of four years in Istanbul children, Nutrition Rep. Intern., 15 (1977) 337. 44 Viikari, J., Akerblom, H.K., Nikkari, T., SeppBnen, A., Uhari, M. et al., Atherosclerosis precursors in Finnish children and adolescents, Part 4 (Serum lipids in newborns, children and adolescents), Acta Paediat. Stand., Suppl. 318 (1985) 103. 45 Green, J.G., Serum cholesterol changes in pregnancy, Amer. J. Obstet. Gynecol., 95 (1966) 387. W.E. and Lin, D.S., Cholesterol 46 Pitkin, R.M., Connor, metabolism and placental transfer in the pregnant rhesus monkey, J. Clin. Invest., 51 (1972) 2584. 47 Dahms, B.B. and Halpin, T.C., Pulmonary arterial lipid deposit in newborn infants receiving intravenous lipid infusion, J. Pediat., 97 (1980) 800. 48 Stary, H.C., Proliferation of arterial cells in atherosclerosis. In: W.D. Wagner and T.B. Clarkson (Eds.), Arterial Mesenchyme and Arteriosclerosis (Adv. Exp. Med. Biol., Vol. 43), Plenum Press, New York, 1974, pp. 59-81. 49 Van Furth, R. and Diesselhoff-Den Dulk, M.M.C., Dual origin of mouse spleen macrophages, J. Exp. Med., 160 (1984) 1273. 50 Spector, W.G. and Lykke, A.W.J., The cellular evolution of inflammatory granulomata, J. Path. Bact., 92 (1966) 163. 51 Mariano, M. and Spector, W.G., The formation and properties of macrophage polykaryons (inflammatory giant cells), J. Path., 113 (1974) 1. 52 Stadecker, M.J. and Wright, J.A., Distribution and kinetics of mononuclear phagocytes in granulomas elicited by eggs of Schistosoma mansoni, Amer. J. Path., 116 (1984) 245. I.Y.R., Role of monocytes 53 Bowden, D.H. and Adamson, and interstitial cells in the generation of alveolar macrophages, Part 1 (Kinetic studies of normal mice), Lab. Invest., 42 (1980) 511. 54 Wisse, E., On the Kupffer cell origin of Kupffer cells, Blood Cells, 6 (1980) 91. P.H. and Volkman, A., Resi55 Sawyer, R.T., Strausbauch, dent macrophage proliferation in mice depleted of blood monocytes by strontium-89, Lab. Invest., 46 (1982) 165. P.H. and Mora56 Volkman, A., Chang, NC., Strausbauch, hart, P.S., Differential effects of chronic monocyte depletion of macrophage populations, Lab. Invest., 49 (1983) 291.

57 De Bakker, J.M., De Wit, A.W., Onderwater,

58 59

60 61

62 63

J.J.M., Ginsel,

L.A. and Daems, W.T., On the origin of peritoneal resident macrophages, Part 1 (DNA synthesis in mouse peritoneal resident macrophages), J. Submicrosc. Cytol., 17 (1985) 133. Still, W.J.S., The topography of cholesterol-induced fatty streaks, Exp. Molec. Path. 20 (1974) 374. Gerrity, R.G., Goss, J.A. and Soby, L., Control of monocyte recruitment by chemotactic factor(s) in lesion-prone areas of swine aorta, Arteriosclerosis, 5 (1985) 55. _ Stary, H.C., Regression of atherosclerosis in primates, Virchow’s Arch. (Path. Anat.), 383 (1979) 117. Zinserling, W.D., Untersuchungen liber Atherosklerose, Part 1 (Uber die Aortaverfettung bei Kindem), Virchow’s Arch. (Path. Anat.), 255 (1925) 677. Aschoff, L., Die Arteriosklerose, Mediz, Klinik, 26 (Suppl. 1) (1930) 1. Stary, H.C., Comparison of the morphology of atherosclerotic lesions in the coronary arteries of man with the morphology of lesions produced and regressed in experimental primates In: M.R. Malinow and V. Blaton (Eds.), Regression of Atherosclerotic Lesions, Plenum Press, New

York, 1984, pp. 235-254. 64 Malinow, M.R., Atherosclerosis: Progression, regression, and resolution, Amer. Heart J., 108 (1984) 1523. 65 Clarkson, T.B., Bond, M.G., Bullock, B.C. and Matzetta, C.A., A study of regression in Maraca mulatta, Part 4 (Changes in coronary arteries from animals with atherosclerosis induced for 19 months and then regressed for 24 or 48 months at plasma cholesterol concentrations of 300 or 200 mg/dl), Exp. Molec. Path., 34 (1981) 345. in 66 Stary, H.C., Evolution and progression of atherosclerosis the coronary arteries of children and adults. In: Atherogenesis and Aging, Springer-Verlag, New York, 1987. 67 Fry, D.L., Responses of the arterial wall to certain physical factors. In: Atherogenesis: Initiating Factors (Ciba Found. Symp. 12), Elsevier, Amsterdam, 1973, pp. 93-125. 68 Nauman, A. and Schmid-Schiinbein, H., A fluid-dynamicist’s and a physiologist’s look at arterial flow and arteriosclerosis. In: G. Schettler et al. (Eds.), Fluid Dynamics as a Localizing Factor for Atherosclerosis, SpringerVerlag, Berlin, Heidelberg, 1983, pp. 9-25.