INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary hypertension (PH) is hemodynamically defined as a mean resting pulmonary arterial pressure (Ppa) above 20 mm Hg at sea level (1). Vascular histopathologic changes associated with PH include medial hypertrophy and various obstructive intimal lesions, such as fibrotic or cellular thickening, thrombi, and plexiform lesions in pulmonary arteries (2). Pulmonary arteries of PH patients usually show some degree of medial hypertrophy and intimal thickening regardless of the etiology of the condition, but thrombotic or plexiform arterial lesions are seen only in certain types of PH (2). Recent advances in cellular and molecular biology have opened new possibilities for examining these histopathologic lesions of vascular remodeling from different angles, which may substantially change the understanding of the pathogenesis of PH. However, careful characterization of the cellular and structural elements of key vascular lesions in PH would be a prerequisite for the application of novel techniques.

A previous study showed that an intratracheal injection of platelet-derived growth factor (PDGF) into rats caused increased DNA syntheses in selected types of cells at different levels of the pulmonary vasculature (3). We hypothesized that some differences may exist in the responsiveness to various stimuli or injuries within the pulmonary vascular system, and that the distribution patterns of pulmonary arteriopathy in different types of PH might give some insight into the pathogenesis of PH. However, published quantitative data on the distribution of pulmonary arteriopathy in different types of PH are very limited (4, 5). This is probably partly due to the difficulty in assessing in vivo vascular changes in routinely processed histologic sections of the lungs that are not fixed with a controlled step of vascular inflation, a technique not very practical for use in a retrospective study of human cases of PPH. Recently, a computer-assisted image analysis method applicable to uninflated lung sections has been established (6). In the present study, we used this method to study the distribution of pulmonary arteriopathy in a large number of documented cases of PH in our pathology file.

Plexiform lesion is a hallmark of obstructive intimal remodeling associated with severe PH. Differentiation of recanalizing thrombi and plexiform lesions from one another can be difficult despite their prognostic implications as reported in a national prospective study of primary pulmonary hypertension (PPH) (9). Some controversies have surrounded the phenotype of the cells in plexiform lesions, and also the presence of plexiform lesions in CTED (2, 10, 11). We therefore examined morphologic manifestations of obstructive intimal lesions in various types of PH and their cellular phenotypes through immunohistochemical methods.

Specific aims of the present study were to: (1) quantitatively examine the degree of obstructive intimal thickening and medial hypertrophy according to the size of arteries in different types of PH; (2) identify any pattern in the distribution of arteriopathy in different types of PH; (3) investigate the differential distribution of plexiform lesions in PPH, Eisenmenger's syndrome (EISEN), and chronic major vessel thromboembolic disease (CTED); and (4) examine the morphologic manifestations of obstructive intimal lesions in PH, with identification of their cellular phenotype.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Case Selection and Review

All of the cases examined in the study were identified from the autopsy or surgical pathology file at the University of California San Diego Medical center from 1990 to 1998. The study was approved by human subject committee at University of California San Diego. Autopsy or surgical pathology reports and medical records were reviewed to verify the clinical diagnosis of PH, underlying conditions, and other pertinent clinical and laboratory data. All available radiologic studies, including plain chest radiographs and chest computed tomographic (CT) scans were also reviewed by the radiologists. Lung histologic sections with hematoxylin and eosin (H&E) and elastin stains were reviewed in each case. Cases were classified as precapillary PPH, EISEN, CTED, pulmonary venoocclusive disease (PVOD), and as miscellaneous for cases associated with fenfluramine-phentermine (fen-phen therapy), scleroderma, and Takayasu's arteritis. Case classification was primarily based on clinical and laboratory data together with radiologic findings, if these were available. The diagnosis of PPH was made after excluding any congenital or acquired cause for PH, but cases associated with systemic lupus erythematosus (SLE) or liver cirrhosis were included in the PPH group. PVOD can be categorized as PPH in a broad sense, but was analyzed separately from the precapillary form of PPH. The clinical diagnosis of PVOD was confirmed by histopathologic examination in each case. All CTED patients had major-vessel thromboemboli documented by cardiac catheterization, pulmonary perfusion scan, pulmonary angiography, and/or by histologic examination at the time of pulmonary thromboendarterectomy (PTE) or at autopsy. Normal control specimens were derived from unused normal donor lungs contralateral to the transplanted lungs.

Morphometry and Definition of Terms

Computer-assisted image analysis for morphometric measurement was performed on 5-µm-thick sections of selected blocks (from one to five sections; average: 2.6 ± 0.2 sections/case) from each case with combined Masson's trichrome-elastic staining. Vessel images were captured and transferred by the Scion image program (Frederick, MD). Arteries were excluded from analysis if they had any of the following features: (1) a longest/shortest diameter ratio > 2.0; (2) an incompletely circular shape; or (3) collapse of more than one quarter of the vessel wall. Thrombotic or plexiform lesions were not included for the measurement of intimal thickening. Measured parameters of arteries included the longest and shortest external diameters, and perimeters of and areas within internal and external elastic lamina and luminal borders. Perimeters and inner areas were computed automatically with the image analyzer program upon meticulous tracing of each lamina or luminal border. Muscle area (between the internal and external elastic laminae) and intimal area (between the luminal margin and internal elastic lamina) were calculated by subtracting the corresponding inner area from the outer area. Vessel data were then corrected for oblique cut and artifactual contraction by mathematical calculation, using previous methods with a minor modification (6). The final data were expressed as external diameter (ED), percent medial thickness (%MT = 2 × MT × 100/ED), and percent intimal thickness (%IT = 2 × IT × 100/ED) as previously defined (12). The external diameters of the arteries immediately adjacent to well-formed plexiform lesions were measured in cases of PPH, EISEN, and CTED.

Data Analysis and Statistical Method

Pulmonary arteries were classified into the following categories of ED: 50 to 100 µm, 101 to 200 µm, 201 to 400 µm, 401 to 600 µm, and 601 to 800 µm. Anatomic distributions of intimal and medial thickenings were analyzed by plotting accumulated data of %IT and %MT against ED in cases of PPH, CTED, and EISEN as three main clinicopathologic groups of PH, with the results compared with those of normal controls. Data from other miscellaneous PH cases were compared with the foregoing prototypes to identify any similarities. EDs of the arteries associated with plexiform lesions were compared in cases of PPH, EISEN, and CTED to find any differences in distribution. All values are presented as mean ± SEM. Statistical analysis was done with software for one-way analysis of variance (ANOVA) and an appropriate post hoc test (Super ANOVA; Abacusconcept, Torrance, CA). One-way ANOVA and Dunnett's one-tailed test were performed to compare the disease groups and controls. One-way ANOVA and the Tukey-Kramer test were used for comparisons among the disease groups.

Immunohistochemistry

Immunohistochemical staining was performed with antibodies to vimentin, -smooth-muscle actin, desmin, clonal designator (CD)31, CD34, and Factor VIII-related antigen on selected paraffin blocks that were routinely processed after formalin fixation. A total of 11 representative cases were used for immunohistochemical study, which included PPH (n = 3), CTED (n = 3), EISEN (n = 1), fen-phen-therapy-related PH (n = 1), and Takayasu's arteritis (n = 1), and glioblastoma multiforme (n = 2). Glomeruloid vascular proliferation in the tumor stroma is one of the diagnostic features of glioblastoma multiforme, a form of high-grade astrocytoma in the brain, and has been thought to be analogous to the plexiform lesion in PH. Because of this, we also studied typical cases of glioblastoma multiforme for comparison. The antibodies, sources, clones, and dilutions used in the study are listed in Table 1. A standard avidin-biotin-peroxidase complex method was used as previously described (13).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Findings

Among 87 cases of PH initially identified from the pathology file at the University of California San Diego Medical Center, 51 cases that had sufficient clinical information and adequate pathologic material were included in the study. A common denominator in all patients was that they presented with sufficiently severe clinical signs and symptoms associated with PH as to which require a serious therapeutic intervention such as lung transplantation and PTE. The pulmonary artery pressures varied among individual cases as well as disease groups. The average Ppa was 57.3 ± 11.9 mm Hg in CTED, 76.2 ± 23.0 mm Hg in EISEN, and 67.8 ± 15.4 mm Hg in PPH. However, there was no statistically significant difference in Ppa between any two groups (p > 0.05). The clinical diagnosis, demographic and hemodynamic data, and other significant findings in each case are summarized in Table 2. A total of 11 cases of PH had pertinent preoperative radiologic studies available for review. Radiologic findings in each disease were not very specific, but chest CT scans generally showed a more prominent mosaic pattern in cases of CTED than in PPH or EISEN, which generally showed a homogeneous attenuation pattern as previously reported (14, 15).

Distribution of Intimal Thickening

The %IT in normal controls was comparable to that in previous studies (Figure 1). The increase of %IT in PPH (n = 15) was observed most prominently in small arteries and arterioles (50 to 200 µm ED), and was significantly greater than in CTED cases (one-way ANOVA and Tukey-Kramer test p < 0.05). The PPH patients also showed a significant increase in %IT within medium-sized muscular arteries (201 to 400 µm), but not in large arteries of ED > 400 µm, as compared with normal controls (one-way ANOVA and Dunnett's one-tailed test, p < 0.05 and p > 0.05, respectively) (Figure 1). In contrast, CTED cases showed a significant increase in %IT in arteries of 401 to 600 µm ED, as well as increases within smaller arteries (one-way ANOVA and Dunnett's one-tailed test, p < 0.05) (Figure 1). There was no significant increase in %IT among the groups of PH cases and controls in arteries of ED > 600 µm (one-way ANOVA and Dunnett's one-tailed test, p > 0.05). The increase in %IT in EISEN, which was much smaller than in PPH or CTED, was not significant at any level of pulmonary arteries as compared with controls (one-way ANOVA and Dunnett's one-tailed test, p > 0.05) (Figure 1).

View larger version (39K): [in this window] [in a new window]   Figure 1.   Intimal thickening in small arteries (< 200 µm ED) was most prominent in cases of PPH (A). Significant intimal thickening in larger arteries (401 to 600 µm ED) was seen only in CTED cases (A). Arteries ranging from 201 to 600 µm ED showed a significant increase in %MT in both PPH and EISEN (B). *p < 0.05 versus control group, by one-way ANOVA and Dunnett's one-tailed test.

The %IT values for cases of PPH, CTED, and EISEN, shown as bar graphs in Figure 1, were plotted linearly according to the arterial size groups, and were overlaid with the linear plotting of %IT from other cases for comparison (Figure 2). PVOD (n = 3) and PH associated with fen-phen therapy (n = 1) showed a PPH pattern characterized by a predominant small-arterial involvement (Figure 2A). The pattern in Takayasu's arteritis (n = 1) was very similar to that seen in CTED (Figure 2B). Scleroderma-associated PH (n = 1) revealed a marked small-arterial obstruction like that seen in cases of PPH, but the overall involvement pattern also showed some similarity to the pattern in CTED (Figure 2C).

View larger version (15K): [in this window] [in a new window]   Figure 2.   The %IT values shown in Figure 1A were plotted linearly in order to demonstrate distribution patterns (i.e., a marked small-arterial predominance in PPH and a less prominent but more widespread involvement of both small and large arteries in CTED). The distribution of intimal thickening in PVOD and fen-phen therapy-induced PH was similar to the pattern in PPH (A). Takayasu's arteritis showed the same pattern as in CTED (B). Scleroderma showed a hybrid pattern of PPH and CTED (C).

Distribution of Plexiform Lesions

The total numbers of well-formed plexiform lesions associated with measurable pulmonary arteries were 69 in cases of PPH, 33 in cases of EISEN, and 18 in cases of CTED. Intimal lesions of equivocal or intermediate form was not included in the measurement. The size of the arteries associated with unequivocal plexiform lesions was significantly smaller in PPH than in EISEN, as shown by the EDs of involved arteries (79.0 ± 6.1 µm and 209 ± 17.6 µm [mean ± SEM], respectively; one-way ANOVA and the Tukey-Kramer test, p < 0.05) (Figure 3). All cases of CTED in our study had angiographically and/or histologically proven thromboemboli in large elastic pulmonary arteries. Five of 22 cases of CTED revealed well-formed plexiform lesions that were indistinguishable from those in PPH or in EISEN. All five cases of CTED showing plexiform lesions in lung sections (Cases 10, 11, 16, 20, and 21) had persistent PH despite successful clot removal by PTE. The average EDs of the involved arteries were intermediate between those in cases of PPH and EISEN (149.5 ± 11.4 µm) (Figure 3).

View larger version (12K): [in this window] [in a new window]   Figure 3.   EDs of pulmonary arteries adjacent to plexiform lesions were significantly smaller in PPH than in EISEN. The arterial size of plexiform lesions in cases of CTED was intermediate between that in PPH and that in EISEN. The values are shown as both dot distributions (A) and bar graphs (B).

Distribution of Medial Hypertrophy

Increases in %MT in PPH and EISEN were significant in medium-sized pulmonary arteries of ED 201 to 400 µm as compared with the %MT in normal controls and CTED (Figure 1). In arteries of 601 to 800 µm ED, a significant increase in %MT as compared with control lungs was found only in EISEN (Figure 1). There was no significant medial thickening in CTED at any level of pulmonary vasculature (Figure 1). Differences in the degree and distribution pattern of medial hypertrophy among each group were not as apparent as differences in intimal thickening. Cases associated with other miscellaneous causes also failed to show any distinguishable pattern (data not shown).

Morphologic Manifestations of Obstructive Intimal Lesions

Eccentric proliferation or fibrosis in the intima usually caused mild and partial luminal obstruction and was found at any level of pulmonary arteries in most PH cases. Concentric intimal proliferation, however, was typically accompanied by nearly complete luminal occlusion, and was found most frequently in small muscular arteries and arterioles (Figure 4A). Thrombotic lesions involving large elastic arteries were only seen in cases of CTED (Figure 4B), but small arterial thrombi were found in many other types of PH (Figure 4C). Some cellular recanalizing thrombi in small arteries and arterioles were difficult to distinguish from plexiform lesions (Figure 4D). Well-formed plexiform lesions (Figure 4E) that were indistinguishable from typical plexiform lesions in PPH, EISEN, or fen-phen therapy-induced PH (Figure 4F) were present in five of the 22 cases of CTED in the study. Demonstrable plexiform lesions were not identified in the lungs of cases of Takayasu's arteritis, PVOD, or scleroderma. The morphologic spectrum of obstructive intimal lesions in various conditions is illustrated in Figure 5.

View larger version (157K): [in this window] [in a new window]   Figure 4.   Obstructive intimal thickening in small arteries and arterioles was most prominent in PPH (A), original magnification: ×100, trichrome elastic stain. A recanalizing thrombus in the large elastic artery shown in B (original magnification: ×20, trichrome elastic stain) was exclusively seen in CTED cases, whereas small arterial thrombi such as shown in C (original magnification: ×200, trichrome elastic stain) were seen in PPH and other types of PH. An example of obstructive intimal remodeling that was difficult to classify as either plexiform lesion or recanalizing thrombus is illustrated in D (original magnification: ×200, trichrome elastic stain). A plexiform lesion found in a CTED case is illustrated in E (original magnification: ×200, H&E stain) and was indistinguishable from a typical plexiform lesion in a case of PPH shown in F (original magnification: ×200, trichrome elastic stain).

View larger version (122K): [in this window] [in a new window]   Figure 5.   Obstructive intimal lesions comprising a morphologic spectrum. (A) Plexiform lesion in EISEN (original magnification: ×40, trichrome elastic stain). (B) A lesion intermediate between a typical plexiform lesion and a recanalizing thrombus in a case of EISEN (original magnification: ×100, trichrome elastic stain). (C and D) Early and mature cellular intimal lesions reminiscent of plexiform lesions in fen-phen therapy-induced PH (original magnification: ×200, trichrome elastic and H&E stain, respectively). (E to G) Intermediate intimal lesions in CTED showing features of both recanalizing thrombi and cellular plexiform lesions (original magnification: ×200, trichrome elastic stain). (H) A well-formed plexiform lesion in CTED (original magnification: ×100, H&E stain).

Cellular Phenotypes of Intimal Lesions

The main cellular constituent of any obstructive intimal lesion consisted of plump or spindle cells reminiscent of immature stromal cells within a granulation tissue. These cells were positive for vimentin and -smooth-muscle actin, but were negative for desmin and endothelial markers CD31, CD34, and Factor VIII (Figures 6A to 6F). Cellular intimal lesions contained variable amounts and shapes of vascular channels that were lined by a single layer of cells staining positively for endothelial markers CD31, CD34, and Factor VIII (Figures 6B, 6D, and 6F). The immunoreactivity of any given vascular lesion to each endothelial marker was quite similar, except that Factor VIII stain tended to have a higher background intensity because of the positive staining of underlying fibrinous materials. Immunoreactivity to various phenotypic markers was remarkably consistent, not only among similar lesions but also among the spectrum of lesions in the various conditions causing PH. Mature and early plexiform lesions exhibited immunostaining patterns almost identical to those described earlier. There was no difference in the immunophenotypes of the cells in the plexiform lesions seen in PPH, EISEN, and CTED. The immunohistochemical staining pattern of the complex vascular lesions in glioblastoma multiforme was identical to that seen in a plexiform lesion described earlier (Figures 6G and 6H).

View larger version (137K): [in this window] [in a new window]   Figure 6.   Most cells in the obstructive intimal lesions in PH were positive for anti--smooth-muscle actin immunostain, as shown in (A, C, and E ). Only a single layer of the cells lining slitlike lumens within the lesions stained positively for endothelial marker CD31, as shown in B, D, and F ). Proliferating vascular lesions in the tumor stroma of glioblastoma multiforme demonstrated the same immunoreactive pattern for actin (G) and CD31 (H ) as the obstructive intimal lesions in PH. Each row is a pair of serial sections from the same case. (A and B) PPH (same lesion shown in Figure 4F). (C and D) CTED (same lesion as shown in Figure 4B). (E and F ) Takayasu's arteritis. (G and H) glioblastoma multiforme.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recent studies of PH have been directed toward the mechanisms of vascular remodeling in this condition, such as complex interactions involving deregulated cell growth, alteration of cell differentiation, and activation of matrix and collagen production (16). Plexiform lesion is one of the most complicated forms of intimal remodeling, and has been the focus of many elaborate cellular and molecular studies of the foregoing issues (19). However, there seemed to be some discrepancy among investigators in defining a given intimal lesion as a plexiform lesion or another type of lesion such as a cellular recanalizing thrombus. We therefore attempted to reexamine these problems with a more basic approach in the context of the whole lung. With this approach, the present study provided basic quantitative data on the distribution of pulmonary arteriopathy in major types of PH, and on phenotypic manifestations of intimal remodeling in chronic PH.

Our study demonstrated that PPH showed prominent obstructive intimal thickening and formation of plexiform lesions primarily at the level of small arteries and arterioles. This finding supports a previous theory of the pathogenesis of PPH, positing endothelial damage within small arteries and arterioles as the initial event, followed by intimal remodeling and luminal obstruction in the damaged arteries (1). The small pulmonary arteries and arterioles in cases of CTED in our study showed significantly less intimal thickening than in PPH. All CTED patients in this study either required lung transplantation for disease refractory even to repeated PTE procedures, or died of operative complications or with PH despite successful removal of large-vessel thrombi. Our results therefore are mainly for severe cases of CTED, probably with more pronounced small-vessel involvement than in typical CTED. Considering such a significant selection bias in the study, it is remarkable that our cases of CTED still showed much less intimal thickening in small arteries and arterioles than did PPH. This may explain a favorable outcome in most CTED patients who undergo PTE procedures. Since our study was done with archival lung sections taken randomly, it was difficult to assess whether any given distal lung section available for review was from an obstructed or an unobstructed area. However, some sections containing both obstructed thrombotic large arteries and adjacent distal small arteries showed a marked obstruction of small-arterial lumina by recanalizing thrombi or fibrotic/cellular intimal thickening. This finding, together with a characteristic mosaic pattern in the CT scan in CTED, suggests uneven small-arterial involvement (possibly more in obstructed and less in unobstructed areas of circulation), as opposed to a rather uniform pattern of small-arterial involvement in PPH. We concluded that the degree and distribution of arteriopathy in the cases of CTED with plexiform lesions still seemed to differ somewhat from the pattern in classic cases of PPH. PTE alone seemed to be ineffective in the CTED patients with plexiform lesions. The degree of intimal thickening in EISEN was significantly less than in PPH or CTED at all levels of pulmonary arteries, despite a long clinical history of disease and high Ppa in our cases.

Plexiform lesions in EISEN were distributed over arteries of a wide range of sizes, as opposed to those in PPH, which were associated mainly with small arteries and arterioles. A previous study, using the size of adjacent bronchi for comparison, also demonstrated the different distribution of plexiform lesions in PPH and EISEN (22). Although CTED has been generally thought to be devoid of plexiform lesions, a previous histologic study of CTED reported a full range of intimal remodeling, including plexiform lesions (11). We also observed well-formed plexiform lesions in five of the 22 cases of CTED in the present study. All of these five CTED patients failed to show a dramatic hemodynamic improvement despite successful PTE, suggesting that plexiform lesions in the distal pulmonary arteries were probably responsible for their persistent PH. Other cases with persistent pulmonary hypertension after PTE (Cases 4, 7, and 19) also showed some intimal proliferative lesions reminiscent of plexiform lesions, though not demonstrable as definite plexiform lesions. Plexiform lesions in EISEN consisted of somewhat more elaborate vascular channels than in PPH or CTED, which instead showed a more prominent cellularity. This suggests a different biologic nature of the plexiform lesions in these conditions, as noted in a previous study (23).

We attempted to use the characteristic distribution patterns of obstructive intimal lesions observed in plexiform PPH, CTED, and EISEN as the prototypes of pulmonary arteriopathy for precapillary small-artery obstruction, large-artery obstruction, and high pulmonary blood flow with increased pressure, respectively. PVOD is a postcapillary obstructive disease of unknown etiology. The distribution pattern of intimal thickening in PVOD was similar, though not identical, to the pattern in plexiform PPH. Diet drug-induced PH has been well documented to have a close clinicopathologic similarity to plexiform PPH (24). Our case of PH with a history of fen-phen therapy, a recently banned appetite-suppressant therapy, also demonstrated a pattern of intimal thickening like that seen in PPH, suggesting a pathogenetic role of endothelial injury in small arteries in diet drug-induced PH. Clinical and pathologic findings in this case have been described in a separate case report (27). A strong association between scleroderma and PH has been well recognized, although the direct pathogenetic mechanism is not clear (28). One case included in the present study showed marked intimal thickening in small arteries and arterioles, comparable to the finding in PPH, but coexisting involvement of larger arteries was not typical of the pattern in PPH, suggesting a hybrid manifestation of PPH and CTED, as seen in the lupus anticoagulant case. Features in PH associated with Takayasu's arteritis were very similar to those in CTED, suggesting large artery obstruction due to vasculitis as the main pathogenetic event.

Despite the lack of distinct differences in the degree and distribution of medial hypertrophy among various types of PH, a relative prominence of medial hypertrophy at all levels of pulmonary arteries in EISEN was noteworthy. Attempts to statistically correlate the pressure and the degree of medial hypertrophy were unsuccessful. However, some cases (especially of EISEN with very high Ppa) certainly gave such an impression. This was an interesting finding, since intimal thickening in EISEN was only meager throughout all levels of pulmonary arteries. It was also interesting that there was no significant medial hypertrophy in CTED at any level of pulmonary artery.

We observed a morphologic spectrum of intimal lesions in PH in terms of cellularity and architecture, with frequent intermediate forms between plexiform and thrombotic lesions that could not be objectively classified as either kind of lesion. A previous study of familial PPH also demonstrated a marked heterogeneity of pathologic lesions, with a frequent coexistence of thrombotic and plexiform lesions, and concluded that these two lesions may not represent distinctly different biologic processes, but may simply be different manifestations of the same process (29). An identical immunohistochemical staining pattern of the two lesions and the presence of intermediate forms in the present study also supported such a conclusion. Histologic and immunohistochemical features of proliferative and obliterative intimal lesions in PH were very similar to those of granulation tissue, since both types of lesions were mainly composed of myofibroblastic cells, with variable amounts of newly formed vascular lumina lined by plump endothelial cells. This suggests a similar nature of these lesions in different locations (i.e., intravascular and extravascular).

Previous ultrastructural studies reported that cells in plexiform lesions were predominantly composed of myofibroblasts, and also what were presumed to be vasoformative reserve cells and smooth-muscle cells (30). The present study also revealed that most abluminal cells in plexiform lesions were positive for -smooth-muscle actin but negative for endothelial markers, supporting a myofibroblastic phenotype of these cells. Another consistent and salient finding was positive staining for endothelial markers (CD31, CD34, and Factor VIII) only in cells lining the vascular channels (well-formed or incipient) within plexiform lesions. Some subtle difference may exist in reactivity to each of the foregoing endothelial markers according to the stage of endothelial differentiation, but it was not apparent to us. More mature plexiform lesions seemed to show better developed, more elaborate vascular channels that were lined by cells positive for endothelial markers. We could not find any reliable markers that could be used for the morphologic diagnosis of plexiform lesions. A recent study proposed an endothelial origin of cells based on positive immunostaining for Factor VIII (10). A subsequent study by the same group of investigators reported a monoclonality of proliferating endothelial cells in primary but not in secondary PH (23). Our result suggests that this issue may need to be revisited after resolution of the controversy surrounding the cell type within plexiform lesions.

In summary, we have provided quantitative morphometric data on pulmonary arteriopathy in documented cases of PPH, EISEN, CTED, and other miscellaneous types of PH. We observed a characteristic distribution pattern for obstructive intimal thickening and plexiform lesions among the major types of PH, which may be useful in categorizing less understood variants of PH. Various proliferative and obliterative intimal lesions in PH consisted mainly of cells showing an immunoprofile of myofibroblasts, with variable endothelial/vascular differentiation. These lesions may represent different manifestations of a similar process, probably influenced by different cytokines and growth factors.