**4. Immunocytochemical typing of subendothelial cells**

The use of antibodies specific to various types of mesenchymal cells has provided additional information regarding the cellular composition of normal and atherosclerotic intima. In addition to cells that expressed smooth muscle α-actin [34-37], immune-inflammatory cells such as macrophages [38-42], lymphocytes [40, 42-47], mast cells [48], and dendritic cells [49] were identified in the aortic intima.

Using a set of cell type-specific cytochemical markers, cells of undiseased and atherosclerotic intima have been characterized (Table 1). The majority of cell population in normal and atherosclerotic intima was found to be represented by cells expressing smooth muscle α-actin. About two-thirds of cells in the muscular-elastic layer expressed smooth muscle α-actin, while in the proteoglycan-rich layer, the proportion of these cells was lower (Table 1). Immuneinflammatory cells (lymphocytes and macrophages) were confined preferentially to the juxtaluminal part of the proteoglycan-rich layer. Their proportion was found to increase in atherosclerotic lesions, reaching 20% of the total cell content.


a : Lesion type according to AHA Council on Atherosclerosis [50] is indicated in the square brackets.

b : The number of cases examined is indicated in the round brackets.

\*: Significant difference from the percent of positively stained cells in grossly normal areas, p<0.05.

ND: Not determined.

**Table 1.** Immunocytochemical identification of cells in human aortic intima

The presence of pericyte-like resident cells in normal and atherosclerotic human aortic intima have been demonstrated [10, 11], using 3G5 antipericyte antibody [51]. The antigen for 3G5 antibody represents O-sialoganglioside of the plasma membrane of brain capillary pericytes [51], which are known as a very stable, slowly renewed cell population [52]. It has been stated that this antigen is typical for quiescent pericytes [51]. Previously, 3G5-positive pericyte-like cells were identified in bovine aorta, in human undiseased intima, and in complicated atherosclerotic plaques with ectopic osteogenesis [10]. These cells account for more than 30% of the total intimal cell population (Table 1). On cross sections, cells bearing 3G5 antigen were found to present only in the uppermost subendothelial layer of the intima. On *en face* prepa‐ rations, 3G5-positive cells with long processes form a network immediately under the endothelium (Figure 4) [1, 11-14]. The subendothelial localization of pericytes suggests that in the aorta and large arteries, 3G5-positive cells perform the functions similar to those of capillary pericytes [1, 11].

**4. Immunocytochemical typing of subendothelial cells**

atherosclerotic lesions, reaching 20% of the total cell content.

2.2±0.4 (3)

6.2±1.2 (4)

5.0±0.9\* (3)

6.2±1.8\* (9)

: The number of cases examined is indicated in the round brackets.

**Table 1.** Immunocytochemical identification of cells in human aortic intima

were identified in the aortic intima.

**Smooth muscle αactin**

> 47.6±2.3 (4)b

> 47.2±3.1 (3)

> 42.2±3.1 (4)

47.0±10. 9 (5)

**Examined area**

208 Muscle Cell and Tissue

Grossly normal [0]a

Initial lesions [I]

Fatty streaks [II]

Fibrolipid plaques [Va]

Fibrotic plaques [Vc]

ND: Not determined.

a

b

The use of antibodies specific to various types of mesenchymal cells has provided additional information regarding the cellular composition of normal and atherosclerotic intima. In addition to cells that expressed smooth muscle α-actin [34-37], immune-inflammatory cells such as macrophages [38-42], lymphocytes [40, 42-47], mast cells [48], and dendritic cells [49]

Using a set of cell type-specific cytochemical markers, cells of undiseased and atherosclerotic intima have been characterized (Table 1). The majority of cell population in normal and atherosclerotic intima was found to be represented by cells expressing smooth muscle α-actin. About two-thirds of cells in the muscular-elastic layer expressed smooth muscle α-actin, while in the proteoglycan-rich layer, the proportion of these cells was lower (Table 1). Immuneinflammatory cells (lymphocytes and macrophages) were confined preferentially to the juxtaluminal part of the proteoglycan-rich layer. Their proportion was found to increase in

**CD45 CD68 3G5 2A7**

3.9±0.4 (5)

6.1±1.4 (4)

13.2±0.8\* (5)

13.1 ±2.3\* (4)

ND ND ND ND ND

: Lesion type according to AHA Council on Atherosclerosis [50] is indicated in the square brackets.

\*: Significant difference from the percent of positively stained cells in grossly normal areas, p<0.05.

The presence of pericyte-like resident cells in normal and atherosclerotic human aortic intima have been demonstrated [10, 11], using 3G5 antipericyte antibody [51]. The antigen for 3G5

**Positively stained cells, %**

0.0±0.0 (3)

1.2±0.3 (3)

3.0±0.7 (3)

(5) 27.0±3.1 (3)

31.3±7.0 (4)

6.3±1.0\* (3)

11.7±2.0\* (8)

5.0±0.7\*

**Inflammatory cells (CD45+CD14)**

> 5.5±1.2 (8)

9.6±1.4\* (6)

13.4±1.5\* (10)

18.7±2.0\* (8)

> 6.4±1.9 (6)

**Resident cells**

97.3±0.6 (8)

90.9±1.2 (6)

89.4±1.1 (10)

84.7±1.4 (8)

94.9±1.6 (6)

**Figure 4.** Positive cells by means of their long processes form a network immediately under the luminal endothelium in undiseased intima. *En face* specimen of the human aorta; immunohistochemistry.

It has been suggested that pericytes contribute to various pathophysiological processes associated with microcirculation dysfunctions, including diabetes, inflammation, wound healing, hypertension, tumor growth [52-54], and also regulate endothelial functions, includ‐ ing endothelial cell proliferation and ion and molecule exchange [53-57]. At the present time, the major predisposition of pericytes is assumed as precursors for other cells of mesenchymal origin, including smooth muscle cells [54], osteoblasts [56-59], chondrocytes [54], and adipo‐ cytes [60]. This is consistent with the earlier concepts that consider subendothelial cells as pluripotent cambial cells [30] and mesenchymal reserve cells [28]. The network formed immediately under the luminal endothelium by pericyte-like cells has been suggested to play a role in immune function of the arterial wall, especially taking into account that pericyte-like cells express HLA-DR antigen [1].

A comparison of the number of pericyte-like cells in atherosclerotic lesions with the number of pericyte-like cells in undiseased intima revealed that in atherosclerotic lesions, the number of the cells possessing 3G5 antigen is much lower than that in undiseased intima (Figure 5).

**Figure 5.** Visualization of 3G5+ cells in normal (undiseased) intima (**A**), in fatty streak (**B**), and in atherosclerotic pla‐ ques (**C**) by means of immunofluorescent analysis.

This could suggest that the number of pericyte-like cells in atherosclerotic lesions decreases. However, the results obtained in experiments with the use of antipericyte antibody 2A7, which recognizes another pericyte antigen, argue with such a suggestion. 2A7+ cells were found to appear in atherosclerotic plaques, while 2A7+ cells are absent in normal intima (Table 1). Anti-2A7 represents an antibody against melanomaassociated high-molecular-weight antigen (HMW-MAA), which is a chondroitin sulfate proteoglycan (also termed as melanoma proteoglycan), which is also present on pericytes in the areas of active angiogenesis (in granulation tissues in healing wounds, synoviitis, etc.) [61]. 2A7 is expressed by "activated" pericytes capable of proliferating [61].

The above-described findings led to a question: are 3G5 and 2A7 pericytic antigens expressed by the same cell or by different cell types? In order to answer this question, a culture of human brain pericytes was used for experiments. It has been found that in human brain pericytes, about 40% of cells express 3G5 antigen and 80% of cells express 2A7 antigen. Simultaneous staining for both antigens revealed 80% of positively stained cells indicating that all cells having 3G5 antigen expressed 2A7 antigen; however, there was a population of pericytes that expressed only 2A7 antigen. As was mentioned above, 2A7 and 3G5 antigens were described on the cells in different functional states: 3G5 antigen is typical for "quiescent" but 2A7 for "activated" pericytes, respectively. Activated 2A7-positive pericytes are capable of intense replication. It can be hypothesized that most of the cultured pericytes are "activated" (as a result of serum stimulation); therefore, only a part of these preserves 3G5 antigen, which is typical for quiescent pericytes.

It is well known that lipid accumulation is one of the most pronounced manifestations of atherosclerosis. Lipid accumulation may influence the expression of pericyte antigens in the subendothelial cells. In fact, a 1.5- to 2-fold increase in the intracellular lipid content induced by modified low-density lipoproteins (modLDL) leads to a decrease in the proportion of 3G5 posilive cells (Figure 6). The total cell number remains unchanged, which indicates that a decrease in the proportion of 3G5+ cells does not result from selective death of these cells but is due to disappearance of 3G5 antigen caused by intracellular lipid accumulation. This occurs in atherosclerotic lesions, where the number of 3G5-positive cells is lower than in normal intima. It can be suggested that the functional state of the pericyte-like cells in the proteoglycanrich layer of the intima is changed upon formation of the atherosclerotic lesions, which is accompanied by alterations in antigenic expression.

**Figure 6.** The effect of intracellular lipid accumulation on the expression of 3G5 pericyte-associated antigen in cultured cells from undiseased human aortic intima. On day 4, cells were washed with sterile PBS and incubated for 24, 72, hand 120 h in Medium 199 with 10% LDS (control), Medium 199 with 10% LDS + 100 µg/ml native LDL (not shown), or Medium 199 with 10% LDS +100 µg/ml modified (desialylated) LDL (modLDL). After incubation, cells were washed with PBS and stained with antibody against 3G5 pericyteassociated antigen. Total cell number and number of positive‐ ly stained cells were calculated in each cell culture in triplicates. The data are representative of the three experiments performed. The data obtained with native LDL were similar to results without LDL and were not shown.
