SYNOVIAL MEMBRANE AND FLUID MORPHOLOGIC ALTERATIONS IN EARLY RHEUMATOID ARTHRITIS : MTCROVASCULAR INJURY AND VIRUS-LIKE PARTICLES H. Ralph Schumacher, Jr. Department of Medicine University of Pennsylvania School of Medicine and Arthritis Section Veterans Administration Hospital Philadelphia, Pennsylvania I9104
The earliest stages of rheumatoid arthritis ( R A ) have received little study, even though this may be the best time to search for mechanisms that initiate joint inflammation. Kulka et a / . in 1955 were impressed by the prominence of microvascular injury in their two earliest cases seen with suspected RA. We also described synovial vascular changes, including erythrocyte extravasation, thromboses, and neutrophil infiltration of venule walls, in our first report of synovitis of less than 1 month duration.' Both studies 1. showed some lining cell proliferation, even in the earliest cases, but a striking paucity of plasma cells. Synovial fluid leukocyte counts were often lower than in chronic RA. Mononuclear cells predominated in synovial effusions in many patients in these series and also in the report of Gatter and Richmond." This report describes a detailed study of the synovial membrane and the synovial fluid cells in eight patients seen during the first 6 weeks of arthritis, who now can be classified as definite or classic RA by the ARA criteria.' These patients are the only ones yet identifiable as RA out of a total of 104 patients seen since 1967 with arthritis of such recent onset. MATERIALS AND METHODS Eight patients initially without definite diagnoses but who now have evolved into classic RA are reported. TABLE1 lists their disease duration at time of synovial biopsy, initial rheumatoid factor titers, synovial fluid findings, and subsequent course. Each patient underwent a needle biopsy of the synovium with the ParkerPearson needle.5 Specimens were immediately placed into Bouin's solution for light microscopy and into half-strength Karnovsky's paraformaldehyde-glutaraldehyde fixative for electron microscopy. The synovial fluids were centrifuged, and the pellets were immersed in the same fixative. After 2 min of fixation, the synovial specimens and the exudate pellets were diced into 1 X l-mm pieces and replaced for a total of 4 hr, at room temperature, in the fixative solution. Specimens were washed in 0.1 M sodium cacodylate buffer at 4" C overnight. They were then postfixed in cold Palade's osmium-veronal for 2 hr, dehydrated in alcohol, placed in a 50% propylene oxide and EponB mixture for 2 hr, and then embedded in Epon 812. Thick sections ( 1 pm) were cut on an LKB-2
39
%
4
38
E
tl
z
P 0
41
Schumacher: Synovial Membrane Alterations
ultramicrotome with a glass knife, stained with toluidine blue, and examined for orientation. Thin sections were cut with a diamond knife and stained with uranyl acetate and lead citrate and then examined on an RCA EMU-3H electron microscope under a 50-kV beam. RESULTS Findings on light microscopy of the synovial biopsies are summarized in TABLE2. Note that some mild lining cell proliferation was apparent in all biopsies. At least occasional diffuse or perivascular lymphocytes were also seen in all specimens, and polymorphonuclear leukocytes (PMN) infiltrated the superficial synovium in six cases. Plasma cells were found in only patient 7, who had arthritis of 1-month duration. FIGUREl a shows a typical mildly
TABLE2 FINDINGS ON SYNOVIAL MEMBRANEBY LIGHTMICROSCOPY Patient 1 2 3 4 5 6 7 8
SLC
Layers 2-4 2-5 3-4 3-4 3-4 1-3 2-6 2-5
PMN 'r
Lymphocytes
2+
2+
+ + + 2+ + 2+ +
2+ 2+
2+ 3+
+
Plasma Cells
+
Vascular Obliteration
+ + + 3+ +
* SLC, synovial lining cells.
t PMN, polymorphonuclear leukocytes.
$
+, Definite but small numbers; 2 + ,
intermediate; 3 + , maximum.
involved area from patient 8 with P M N scattered throughout the center of the villus. FIGURE l b from patient 4 illustrates the type of vascular obliteration often seen. Vessels appeared to be plugged by inflammatory cells and also by organized thrombi. In FIGURE 2, we see more intense inflammation, including PMN and lymphocytes with narrow vessel lumens. Lumen narrowing possibly occurs merely because of temporary slower perfusion of this area. In this patient (no. 7 ) , other areas exhibited actual infiltration of vessel walls with PMN. This finding was the only suggestion of true vasculitis and is illustrated elsewhere (Reference 2, Figure 4). All changes were focal; some areas in several biopsies appeared perfectiy normal. Vascular congestion and edema were often noted. The synovial membrane electron microscopic findings are given in TABLE 3. There was no definite pattern for the type of synovial lining cell (SLC) that seemed to predominate. Type-A (phagocytic) and intermediate cells when present possessed long cytoplasmic processes and often contained finely granular
42
Annals New York Academy of Sciences
FIGUREla. Synovial villus from patient 8. There is dramatic vascular congestion, 2-3 layers of synovial lining cells (SLC), and a few scattered lymphocytes and polymorphonuclear leukocytes (arrow). Light microscopy; hematoxylin eosin stain. x 360.
material in their vacuoles. No virus-like particles were seen in SLCs (FIGURE 3). Occasional plasma cells were found by electron microscopy in two patients. Interstitial necrotic debris was common. I n all except one patient, vascular alterations were prominent. Gaps were noted between endothelial cells, and in these areas in particular, the endothelial cytoplasm displayed many flaps and processes that when cut on cross section might be superficially confused with 4 ) . Alterations involved both the superficial fenestrated viral particles (FIGURE venules and deeper thick-walled venules. Arteriolar disease was not observed. Vascular endothelium was frequently prominent and often showed rodlike Weibel-Palade bodies, as are also seen in normal vessels.x In the past, these bodies had been thought to be possible microorganisms.3fi There were also many filaments, occasional lipids, dense bodies, free ribosomes, cisternae of rough endoplasmic reticulum, pinocytic vesicles, mitochondria, multivesicular bodies, and normal appearing nuclei. Some endothelial cells had lost the normal cytoplasmic density and appeared to be degenerating. Four patients had some vessel lumens obliterated with platelet thrombi (FIGURE 5 ) . A single clump of paramyxovirus-like structures as seen in systemic lupus erythematosus was found in the vascular endothelium of patient 7.
Schumacher : Synovial Membrane Alterations
43
FIGUREI b. Superficial synovial vessel lumens plugged by inflammatory cells (arrows) from patient 4; light microscopy; hematoxylin and eosin stain. ~ 3 6 0 .
Endothelial basement membranes were often multilaminated. Pericytes and connective tissue components were frequently widely separated by low-densityareas, which suggested edema. There were extravasated erythrocytes and perivascular lymphocytes and PMN. PMN both perivascularly and within vascular lumens were degranulated, as previously described.!' I n four patients, particles were present that deserve consideration as to their possible viral nature. FIGURE 6 illustrates particles that appear to be budding from the vascular endothelium of patient 2. The particles in this patient are about 100-140 nm in diameter. FIGURE 7 from patient 6 also suggests a budding form: here, the smaller 90-nm budding structure seems to protrude into a vacuole. Each particle has a dense core surrounded by lucent areas and unit membranes and thus might be considered a C-type particle.lf1Il 1 Patients 6 and 8 also exhibited 80-140-nm particles in the perinuclear space without any definite budding. They did not have distinct central cores and 8 ) . Pacould also have been invaginations of endoplasmic reticulum (FIGURE tient 5 had 80-140-nm structures present in the vascular lumen near the endothelium, with a few central ribosome-like granules that are certainly not 9a). diagnostic but are identical in appearance to the arenaviruses * ? (FIGURE Many generally round particles with similar granules were also observed adjacent 9b). to a pericyte in patient 2 (FIGURE
C
+
+
PC t
+
2+
+
Surface Fibrin
+ ++ +
Viruslike Particles
t PC, plasma cells. I See footnote to TABLE 2 for definition of symbols.
* SLC, synovial lining cells; A, B, and C refer to types of lining cells.‘
7 8
5 6
4
2+
+ f
B
+ 2+ + 2+ 3+ 2+ 2+ + + + ++ 2 ++ 2+
+ + 3+
1
2 3
A
Patient
SLC 8
2+
+ 2+ ++ + 2+
Endothelial Gaps
+ + +
+ + ++ +
Dense Deposits in Vessel Walls Necrosis
TABLE3 ELECTRONMICROSCOPIC FINDINGS IN SYNOVIAL MEMBRANES
+++
+
Vascular Obliteration
+ +
Fibrin in Vessel Wall
s
2
2. 0
v,
?n
Y
Ea
2
P?
.c
5
E”
J
B
Schumacher : Synovial Membrane Alterations
45
FIGURE 2. Intensely inflamed synovium in patient 7. There are polymorphonuclear leukocytes, lymphocytes, and, in this case only, occasional plasma cells. Vessel lumens are narrowed. Light microscopy; hematoxylin and eosin stain. x 360.
TABLE 4 ELECTRONMICROSCOPY OF SYNOVIAL FLUIDCELLS Phagocytic Large Mononuclear Cells
Synthetic LM
Phagocytized Fine Granular Patient Cells and Debris Material I
2 3 4
5 6 I 8
+ + 2+
+ 31-
+ 2+ 2+ 2 f
%:
PMN Gray Globular Fine Granular Inclusions Inclusions
++ + + + 2t
* See footnote to TABLE 2 for definition of symbols.
+++ ++
+1-
46
Annals New York Academy of Sciences
FIGURE 3. Synovial lining cells (SLC) and superficial capillary of patient 5 with synovitis of 19 days’ duration. The most superficial SLCs and some of the deeper ones have prominent cytoplasmic processes, vacuoles, and dense bodies. There are varying amounts of rough endoplasmic reticulum (RER). These cells are type A (A) or intermediate (I). One type-B cell ( B ) with profuse RER is seen. V, small vessel without any identifiable pathologic change. x 6400.
Schumacher : Synovial Membrane Alterations
47
FIGURE 4. Synovial venule from patient 8 with gap (arrow) between two endothelial cells ( E ) . Many endothelial flaps and processes occur at the site of the endothelial junction opening. Note the loss of normal cytoplasmic density of the lower endothelial cell. This is probably a sign of cell degeneration. RBC, erythrocyte in the lumen; P, pericyte; BM, multilaminated basement membrane; electron micrograph. x 1 1,000.
48
Annals New York Academy of Sciences
FIGURE 5 . Platelets (PL) obliterate the lumen of a superficial synovial vessel from patient 6. G, gap between endothelial cells ( E ) ; BM, basement membrane; RBC, erythrocytes extravasated and lying in the vessel wall. The electron density of hemoglobin is not easily distinguishable from small dense deposits (D) of what may be other proteins in the vessel wall. P, pericytes; electron micrograph. x 12,800.
Schumacher: Synovial Membrane Alterations
49
FIGURE 6a. Virus-like particles (arrows) budding from the synovial vascular endothelium ( E ) of patient 2. Note that in this Figure, the endothelium is thin walled with a fenestration ( F ) . BM, basement membrane. Other particles free in the vessel lumen can be cell debris. Electron micrograph. x 48,400.
50
Annals New York Academy of Sciences
FIGURE 6b. See legend to FIGURE 6a. X 88,000.
In both patients 2 and 6, round particles with some central density were seen in vessel walls surrounded by electron-dense deposits (FIGURES 10a & b ) . Other cell debris that did not especially resemble viruses was also seen between endothelium and pericytes. Patients 4 and 6 also had occasional small dense deposits, without any identified particles between endothelium and pericytes. One other structure that merits consideration as possibly being virus related is shown in FIGURE 11. Here, tubules approximately 200 A in diameter lie in a parallel array in a nipple-like protuberance of a perivascular cell. This configuration is reminiscent of the nucleocapsids illustrated in budding virions of
Schumacher: Synovial Membrane Alterations
51
canine distemper virus l 1 and measles virus,I2 although the tubules are more densely packed in our patient. The known viral nucleocapsids have had smaller diameters of 100-150 A. Electron microscopy of the synovial fluid cells (TABLE 4) revealed no lymphocytes or other cells with paramyxovirus-like inclusions. PMN often contained gray globular bodies but very rarely any vacuoles of similar size with finely granular material (FIGUREI 2 a ) . F a r more evidence of phagocytosis was present in the large mononuclear (LM) cells; the LM cells of four patients contained large vacuoles filled with finely granular material, and five had phagocytized cell debris (FIGURE 12b). Four of the latter had ingested whole cells and were thus comparable to “Reiter’s” cells. No particles were especially suggestive of viruses or other organisms.
FIGURE 7. Ninety-nanometer particle (arrow) budding into an apparent vacuole in patient 6. PV, pinocytic vesicles; M, mitochondrion. The apparent particle in a small cisterna of rough endoplasmic reticulum (ER) may be an invagination of the ER; electron micrograph. x 32,000.
52
Annals New York Academy of Sciences
FIGURE 8. Perivascular cell with degenerating nucleus ( N ) showing loss of the nuclear envelope at the top. In this nucleus, as in many degenerating synovial nuclei in RA, gout, and other diseases, some of the chromatin appears as short tubules (arrows) that resemble the tubules in the tubuloreticular structures of systemic lupus erythematosus. Note also a variety of multivesicular bodies (MVB), unidentified round particles that may be cell debris, and particles in the perinuclear space (double arrows); electron micrograph. x 25,600.
Schumacher: Synovial Membrane Alterations
53
DISCUSSION In this review of an experience now expanded to include eight cases of rheumatoid arthritis seen within the first 6 weeks of disease, emphasis should be placed on the prominence of the microvascular changes. Microvascular injury is easily documented, but mechanisms of injury and the role of microvascular changes in perpetuation of arthritis are more difficult. All microvascular changes reported here (except for the budding virus-like particles) have also been observed in other acute inflammatory diseases, including transient undiagnosed synovitis.2 Microvascular obliteration, dense deposits, or other phenomena do not seem to permit any conclusions with respect to prognosis. However, it is of interest that those rheumatoid patients in this series who presented evidence of possible virus-like particles have had generally milder disease. All patients fit current criteria for RA and have as yet developed no excluding phenomena. Nevertheless, we should note that what is now known as RA may still include a wide variety of diseases with different initiating and perpetuating mechanisms, in addition to different prognoses. If microorganisms can initiate RA, one must suspect that many different organisms may be involved. Viral infection might cause only some “RA.” Because the group in this study was selected on the basis of relatively acute onset, it certainly is not representative of all RA. None of the patients who were latex negative when first examined have subsequently developed rheumatoid factor. We have studied the microvasculature to look for factors that might produce 10a and the initial injury. The virus-like particles are one possibility. FIGURES 10b from this series show virus-like particles embedded in electron-dense material in synovial vessel walls, which raises the intriguing possibility that viral antigen-antibody complexes produce the vascular injury. Such deposits are still far too few to do more than encourage continued search for similar areas in other patients. Virus particle-antiviral antibody complexes have been implicated in glomerular injury in lymphocytic choriomeningitis viral infection 34 and in the glomerulonephritis in murine leukemia.35 At present, we have no evidence of a direct cytopathic effect of any of the virus-like particles on synovial cells. This lack of evidence is in contrast to our work with hepatitis, where suspected viral particles were frequently associated with necrosis of involved ~ e l l s . 4 ~ Observation of occasional dense deposits in vessel walls without any viruslike particles also means that the presence of various other antigen-antibody complexes is still a possibility. Vessel lumens obliterated by platelets and packed with inflammatory cells did not invariably show any changes in the vessel wall. Thus, intraluminal processes might initiate some vascular disease. Other electron microscopic studies in RA have all been performed in wellestablished chronic disease and have rarely mentioned virus-like particle^.^^-^^ The tubular paramyxovirus-like structures so common in systemic lupus erythematosus are rare in RA.lX3143 Helder et d . 2 ’ found structures that suggest paramyxovirus nucleocapsids in only one of 25 RA synovial membranes. Gyorkey et d.lYreported such inclusions in each of four rheumatoid synovial membranes studied but did not describe their criteria for diagnosis or give ~ details about localization of the possible nucleocapsids. Neumark et ~ 1 . ’ examined 30 RA synovia from synovectomies and 17 noninflamed control membranes and found a variety of round particles 600 to 2500 A in diameter. Most were in the cytoplasm of vascular endothelium, macrophages, and other connec-
54
Annals New York Academy of Sciences
FIGURE 9a. Arenavirus-like particles (arrows) with unit membranes and ribosomelike granules. These particles could also not be distinguished from unusual cell processes in this endothelial cell adjacent to a gap ( G ) . 1,vascular lumen; patient 5; electron micrograph. x 49,600.
tive tissue cells. Convincing budding was not shown. Some clumps or isolated 160-1 80-Atubules in nuclei were described as paramyxovirus-like. More recent work by Neumark and FarkasZ3also suggests budding of virus-like particles from cell membranes, but cell membrane was not identifiable at the sites of suspected budding. Serre et aLZOreported but did not illustrate suspected viral nucleocapsids, 18 pm in diameter, in supernatants from tissue cultures of RA synovial membranes. have shown very slightly curved tubular 240-A diameter Helder et dZ4 structures not only in the cytoplasm of histiocytes in RA but also in other tissue
Schumacher : Synovial Membrane Alterations
55
reactions. These structures appeared identical to those found in patient 5 of our series and are not believed to be viral; they are, instead, considered similar to the membranous complexes seen, for example, in histiocytosis X.25 Helder et al. could find no particles like those reported by Neumark. Nuclear bodies in endothelial and perivascular cells have been described in rheumatoid synoviurn.l$ These authors recognized that none of them were specific for virus infection, but they did suggest a possible viral origin. We have not seen any nuclear material suggestive of viruses. Nucleoli13 might account for some of the images shown. Other tubules and filaments in crystallike arrays are similar to structures seen in the cytoplasm in many cells and are of unknown importance.:{h Most authors, including Neumark et al.,'2 have noted that it is difficult to distinguish viral particles from altered cell components, and this point deserves continued emphasis. Although electron micrographs can suggest the presence of viruses, culture will be required to prove viral invasion. Of course, even the culture of a virus does not establish its involvement in the pathologic process under consideration.
FIGURE 9b. Similar particles appearing to have arisen from a perivascular cell in patient 2; electron micrograph. X70,400.
56
Annals New York Academy of Sciences
FIGURE 10a. Dense deposit with virus-like particles beneath the endothelium (E) in patient 2. L, lumen; electron micrograph. x 70,400.
The rarity in RA of the paramyxovirus-like or tubuloreticular structures so typical of IupuslD seems to be a notable difference between these diseases. Recent cytochemical studies have suggested that the tubuloreticular structures may principally contain phospholipid and acid glycoprotein and thus most likely are not viral.51 Some tubular structures seen in our electron microscopic studies seem to be breakdown products of nuclear chromatin unrelated to viral 8 ) . Whether any of them have been considered virus-like infection (FIGURE tubules in the past is unknown. Definite identification of particles as viruses in electron microscopy thin sections is difficult, because a variety of cell components and still unidentified particles can superfically resemble virus particles.15 Most often, nonviral material resembles C-type viral particles, which have a dense core with a surrounding unit membrane. More intense osmium staining of the core or membrane, as 6 and 7 , tends to favor the presence of actual virus. illustrated in FIGURES Sources of possible confusion include transverse sections of collagen fibers, microvilli, and other cell processes (FIGURE4 ) , nuclear pores; annulate lamellae; vesicles in multivesicular bodies (FIGURE 8) ; secretion granules; pinocytic vesicles with dense inclusions, dense bodies, namely, lysosomes and platelet
Schumacher: Synovial Membrane Alterations
57
granules; mycoplasma elementary bodies; some hyperplastic endoplasmic re7 ) ; hyperplastic Golgi apparatus; and large nuclear granules. ticulum (FIGURE Dense bodies usually do not have the same wide space between the dense core and the membrane as viral C-type particles do. Membrane spikes observed on virus membranes at very high magnification also aid differentiation. Golgi apparatus, nuclear envelope, and pinocytic vesicles have all been described as contributing to the multivesicular bodies (MVB) .*';*li Their ubiquity suggests that they are not viral material, but it should not be considered proven that viruses cannot ever be components of MVB." The synovial fluid cell studies in our eight patients revealed no virus-like particles; this result is consistent with the hypothesis that if an infectious agent is involved, the initial insult occurs in the synovium. The dramatic amounts of cell debris present offer potential confusion with microorganisms, but
FIGURE lob. Particles, some of which could be cell debris in a loose appearing granular deposit ( D ) outside endothelium ( E ) in patient 6. Note also that the PMN in the lumen is almost completely degranulated; electron micrograph. x 11,200.
58
Annals New York Academy of Sciences
FIGURE 11. Nipple-like protuberance of a perivascular ribosome-rich cell of patient 2. Densely packed 200-A tubules lie in parallel array (arrows); electron micrograph. x 70,400.
Mitrovic et al.52 demonstrate how all these inclusions can be reproduced by phagocytosis of cell debris. As we have reported previously,Z PMNs in early RA synovial fluids do not seem to have any more than very rare electron microscopic evidence of phagosomes of the size described as cytoplasmic inclusions in “RA cells.’’53 The structures most consistent in size with the cytoplasmic inclusions of light and fluorescence microscopy appear to be the gray globular bodies described by Zucker-Franklin 4 3 that appear more likely to be lipid.5’ Granular material that resembles the complexes prepared by Zucker-Franklin is found much more commonly in the large mononuclear cells, some of which may be type-A SLC released into the joint space or monocyte-derived macrophages. Because of our hope than an infectious initiating agent might be demon-
Schumacher : Synovial Membrane Alterations
59
strable very early in disease, we have continued tissue culture studies since our 1971 report 26 and still can provide no evidence of viral replication. Slow viruses may be agents in such diseases as R A and are notoriously difficult to culture.55 Reasons for continued search for a virus or other infection that initiates R A include clinical observations of chronic polyarthritis that can follow presumed viral upper respiratory infection,27 experimental viral infection that produces chronic synovitis,23 the resistance of rheumatoid synovial cells to certain virus infections (reported but not yet confirmed) and the ability to transfer this resistance to rabbit synovial cells,zg and spontaneous chronic arthritis in chickens attributed to reovirus.:'O, 31 T h e intriguing association of ownership of pets, especially dogs, that antedates the onset of RA suggests
FIGURE 12a. Synovial fluid cells from patient 2. Lymphocytes at the bottom have prominent cytoplasm but no virus-like inclusions. Note the uropod (U).The three PMN above show structures like the gray globular (GG) bodies of Zucker-Franklin 4' that are very homogeneous and appear to be lipid. There are no comparably sized vacuoles with granular material. The small clearer areas at the arrows seem more like sites of degranulation than phagocytic vacuoles; electron micrograph. x 6400.
60
Annals New York Academy of Sciences
that an animal virus may cause the human disease. We have been studying a rheumatoid-like polyarthritis in dogs and have found tubuloreticular structures and crystalline arrays of 2 2 0 4 tubules but no particles comparable to those shown here in human RA.
FIGURE12b. Synovial fluid cells from patient 6. Large mononuclear cells contain vacuoles (V) with fine granular material. One cell has phagocytized a necrotic PMN farrows); electron micrograph. x 6400.
This work so far provides no clues to possible perpetuating mechanisms but does suggest some speculation. If a virus could arrive via the circulation to initiate the synovitis, some yet undetermined factor might determine whether viral antigen is sequestered in the joint to provide an ongoing antigenic stimulus. Rubella virus antigen, for example, has been found in chondrocytes of con-
Schumacher : Synovial Membrane Alterations
61
genitally infected rabbits.4* Synovial and cartilage cells can be infected by viruses in ~ i t r - 0 Organisms .~~ may well not be morphologically identifiable, even though they still provide antigen. In known infections, as with mycoplasma hyorhinis infection of swine,32 it is important to note that organisms may not be identifiable by electron microscopy, despite the fact that they can still be ~ shown that antigen in cultured from the joint. The work of Cooke et U Z . ~has rabbits sensitized to bovine serum albumin (BSA) and then given intraarticular BSA is sequestered in cartilage and other connective tissue. Some initial local inflammatory reaction seems to be necessary in a yet undefined way to allow the complexes to enter and be stored in the collagenous tissues.47 Cell debris, and therefore possibly viral antigen, is scattered throughout the synovial collagenous tissue in our electron micrographs. SUMMARY Eight patients have been studied during the first 6 weeks of rheumatoid synovitis. All of them exhibited microvascular injury, which was manifested by gaps between endothelial cells, vascular occlusion, erythrocyte extravasation, or endothelial cell injury. In four patients, a variety of virus-like particles were found associated with the endothelium or perivascular cells. In two cases, particles were seen in electron-dense deposits in vessel walls. Lymphocytes and P M N infiltrated the synovial membranes, but plasma cells were uncommon. Evidence of phagocytosis was prominent in synovial lining cells and other large mononuclear cells, but not in PMN. These observations are consistent with injury to synovium and, specifically, synovial vessels as an early stage in RA synovitis. The virus-like particles require further investigation, because nonviral cell components remain very difficult to distinguish in electron microscopy tissue sections. REFERENCES
J. P., D. BOCKING,M . W. ROPES& W. BAUER.1955. Early joints lesions 1. KULKA, of rheumatoid arthritis. Arch. Pathol. 59: 129-150. 2. SCHUMACHER, H. R. & R. C. KITRIWU.1974. Synovitis of recent onset. A clinicopathologic study during the first month of disease. Arthritis Rheumat. 15: 465-485. 3. GATTER,R. A. & J . D. RICHMOND. 1973. The predominance of synovial lymphocytes in early rheumatoid arthritis. Arthritis Rheumat. 1 6 544 (abs.). S. COBB,R. JACOX & R. A. JESSAR.1958. Revi4. ROPES,M. W., G. A. BENNETT, sion of diagnostic criteria for rheumatoid arthritis. Bull. Rheumat. Diseases 9: 175, 176. 5. SCHUMACHER, H. R. & J. P. KULKA.1974. Needle biopsy of the synovial membrane. Experience with the Parker-Pearson technic. N. Engl. J. Med. 286: 41 6-4 19. 6. KARNOVSKY, M. J. 1965. A formaldehyde-glutaraldehyde fixative of high osrnolality for use in electron microscopy. J. Cell Biol. 27: 441 (abs.). 7. BARLAND, P., A. B. NOVIKOFF & D. HAMERMAN. 1962. Electron miscroscopy of the human synovial membrane. J . Cell Biol. 14: 207-220. H. R. 1969. The microvasculature of the synovial membrane of 8. SCHUMACHER, the monkey. Ultrastructural studies. Arthritis Rheumat. 12: 387-404. H. R. & C. A. AGUDELO.1972. Intravascular degranulation of 9. SCHUMACHER, neutrophils: an important factor in inflammation? Science 175: 1139, 1140.
62
Annals New York Academy of Sciences
W. 1960. The detection and study of tumor viruses with the electron 10. BERNHARD,
microscope. Cancer Res. 20: 712-727. 1 1 . DALTON,A. J. 1972. Further analysis of the detailed structure of type B and C particles. J. Nat. Cancer Inst. 48: 1095-1097. 12. DALTON,A. J. 1973. The arenaviruses. In Ultrastructure of Animal Viruses and Bacteriophages. A. J. Dalton & F. Haguenau, Eds. : 323. Academic Press, Inc. New York, N.Y. D. W. 1966. The Cell. : 26. W. B. Saunders Company. Philadelphia, 13. FAWCETT, Pa. 1970. Nuclear bodies in rheumatoid synovium. T. & K. FARKAS. 14. NEUMARK, Ann. Rheumat. Diseases 29: 653-659. 15. HAGUENAU, F. 1973. “Virus-like’’ particles as obzerved with the electron microscope. I n Ultrastructure of Animal Viruses and Bacteriophages. A. J. Dalton & F. Haguenau, Eds. : 391-396. Academic Press, Inc. New York, N.Y. 16. MATIN, B. J. & S. S. SPICER.1973. Multivesicular bodies and related structures of the syncytiotrophoblast of human term placenta. Anat. Rec. 175: 15-34. W. & A. JASINSKI.1970. The formation of multivesicular bodies 17. KILKARSKI, from the nuclear envelope. J. Cell Biol. 45: 205-21 I . 18. SCHUMACHER, H. R. 1970. Ultrastructure of the synovial microvasculature in rheumatoid arthritis and systemic lupus erythematosus. I n Proceedings of the Electron Microscopy Society of America. C. I. Arceneaux, Ed. : 240. Claitor’s Publishing Division. New Orleans, La. K. W. MIN & P. GYORKEY.1972. A morphologic F., J. G. SINKOVICS. 19. GYORKEY, study on the occurrence and distribution of structures resembling viral nucleocapsids in collagen diseases. Amer. J. Med. 53: 148-158. 20. SERRE,H., J. MANDIN, C. VAGO,J. SANG& J . CLOT. 1972. Recherche et etude de virus dans la synoviale au cours de la polyarthrite chronique rhumatismale. Rev. Rhumat. 39: 595-599. 21. SCHUMACHER, H. R. 1973. Needle synovial biopsy: evidence for infectious agents including viruses as possible etiologies in arthritis. Excerpta Med. 299: 27 (abs.). T., I. HOLLOS& K. FARKAS. 197 I . Electron microscopical investiga22. NEUMARK, tion of virus-like particles in rheumatoid synovial membrane. Vnitini Lekaistri. 17: 625-630. 23. NEUMARK,T. & K. FARKAS.1973. Ultrastructural aspects of lymphoreticular cells in rheumatoid synovium. Ann. Rheumat. Diseases 32: 524-530. 24. HELDER,A. W., T. M. FELTKAMP-VROOM & R. L. F. NIENHUIS.1973. Electron and light microscopical observations and serological findings in rheumatoid arthritis. Ann. Rheumat. Diseases 32: 5 15-523. 25. BASSET,F., J. ESCAIG& A. LECROM.1972. A cytoplasmic membranous complex in histiocytosis X. Cancer 29: 1380-1386. 26. SCHUMACHER, H. R., S. L. KATZ& R. B. KUNDSIN.197 I . The continued search for viral and other infectious agents in the systemic rheumatic diseases: study of early lesions. In Atypical Virus Infections. C. L. Christian, Ed. : 18-23. The Arthritis Foundation. New York, N.Y. 27. HOLLANDER, J. L., E. M. BROWN,R. A. JESSAR,K. HUMMELER & W. HENLE. 1962. Studies on the relationship of virus infections to early or acute rheumatoid arthritis. Arch. Interamer. Rheumatol. 5: 137-157. 28. WEBB, R. W., R. BLUESTONE,L. S. GOLDBERG, S. D. DOUGLAS & C. PEARSON. 1974. Experimental viral arthritis induced with herpes simplex. Arthritis Rheumat. 1 6 241-250. 29. SMITH,C., D. HAMERMAN, R. JANIS& E. HABERMAN.1974. Virus resistance transferred from human rheumatoid cells to rabbit synovial cells. Ann. Rheumat. Diseases 33: 173-179. 30. GLASS,S. E., S. A. NAQI, C. F. HALL & E. M. KERR. 1973. Isolation and characterization of a virus associated with arthritis of chickens. Avian Diseases 17: 415-424.
Schumacher : Synovial Membrane Alterations
63
E. R., M. H. FRIEDMAN & N . 0. OLSON. 1972. Electron microscopic 31. WALKER, study of an avian reovirus that caused arthritis. J . Ultrastruct. Res. 41: 67-69. J. R. & R. F. Ross. 1969. Fine structure of synovial membrane in 32. DUNCAN, mycoplasma hyorhinis arthritis of swine. Amer. J. Pathol. 57: 171-181. H. R. 1970. Tubular paramyxovirus-like structures in synovial 33. SCHUMACHER, vascular endothelium. Ann. Rheumat. Diseases 29: 445-447. M. 9. A. & F. J . DIXON.1970. Pathogenesis of chronic disease 34. OLDSTONE, associated with persistent lymphocytic choriomeningitis viral infection. 11. Relationship of the anti-lymphocytic choriomeningitis immune response to tissue injury in chronic lymphocytic choriomeningitis disease. J. Exp. Med. 131: 1-19. 35. PASCAL,R. R., M. N . Koss & R. L. KASSEL. 1973. Glomerulonephritis associated with immune complex deposits and viral particles in spontaneous murine leukemia. An electron microscopic study with immunofluorescence. Lab. Invest. 29: 159-165. T. C. & D. G. RAYNS. 1966. Further electron microscopic studies of 36. HIGHTON, the human synovium. Proc. Univ. Otago Med. School 44: 37-39. & D. HAMERMAN. 1964. Fine structure of the 37. BARLAND,P., A. 9. NOVIKOFF rheumatoid synovial membrane with special reference to lysosomes. Amer. J. Pathol. 44: 853-866. 38. GHADIALLY, R. N. & S. ROY. 1967. Ultrastructure of synovial membrane in rheumatoid arthritis. Ann. Rheumat. Diseases 26: 426-443. 39. NORTON,W. L. & M. ZIFF. 1966. Electron microscopic observations on the rheumatoid synovial membrane. Arthritis Rheumat. 9: 589-609. 1971. UltraJ., A. VALLE,L. DIAZ-FLORES & J. M. ARAMBURU. 40. BORRACHERO, estructura de la sinovitis reumatoide. Rhumatologie 1: 277-285. H. J. C., H. M. LAIRD,N. G. WRIGHT & H. THOMPSON.1971. 41. CORNWELL, Ultrastructural features of canine distemper virus infection in a dog kidney cell line. J . Gen. Virol. 12: 281-292. 42. NAKAI,T., F. L. SHAND& A. F. HOWATSON.1969. Development of measles virus in vitro. Virology 38: 50-67. 43. ZUCKER-FRANKLIN, D. 1966. The phagosomes in rheumatoid synovial fluid leucocytes. A light, fluorescence and electron microscopic study. Arthritis Rheumat. 9: 24-36. 44. LONDON, W. T., D. A. FUCCILLO, 9. ANDERSON & J . L. SEVER.1970. Concentration of rubella virus antigen in chondrocytes of congenitally infected rabbits. Nature (London) 226: 172, 173. 45. CRAIG,C. P. & R. B. GREER. 1971. Virus infections in rabbit articular cells. Clin. Res. 19: 439. T. D., E. R. Hum, M. ZIFF & H. E. JASIN. 1972. The pathogenesis of 46. COOKE, chronic inflammation in experimental antigen-induced arthritis. 11. Preferential localization of antigen-antibody complexes to collagenous tissues. J. Exp. Med. 135: 323-338. 47. COOKE,T. D. & S. M. RICHER. 1974. Pathogenesis of chronic inflammation in experimental antigen-induced arthritis. 111. Influence of immune state and local inflammatory response on early localization of antigen to articular collagenous tissues. 1. Rheumat. 1: 143-152. H. R. & E. P. GALL.1974. Arthritis in acute hepatitis and chronic 48. SCHUMACHER, active hepatitis. Pathology of the synovial membrane with evidence for the presence of Australia antigen in synovial membranes. Amer. J. Med. 57: 655-664. N. L., N . DITCHEK, J. POILEY& I. N. KIEM. 1974. Pets and 49. GOTTLIEB, rheumatoid arthritis. Arthritis Rheumat. 17: 229-234. C. D., H. A. ALLEN,R. E. HALLIWELL & H. R. SCHUMACHER. 1974. SO. NEWTON, Clinico-pathologic conference. J. Amer. Vet. Med. Ass. 165: 459-464. 51. SCHAFF, Z., D. BARRY & P. M. GRIMLEY.1973. Cytochemistry of tubuloreticular structures in lymphocytes from patients with systemic lupus erythematosus
64
52.
53. 54.
55.
Annals New York Academy of Sciences and in cultured human lymphoid cells. Comparison to a pararnyxovirus. Lab. Invest. 29: 577-586. MITROVIC,D., J. BARBOT,M. F. KAHN,D. Hroco & S. DE SBZE. 1969. Reproduction experimentale des inclusions gtantes dans les macrophages pulmonaires du lapin. Pathol. Biol. Sern. Hop. 17: 53-58. HOLLANDER, J . L., D. J. MCCARTY,G. ASTORGA& E. CASTRO-MURILLO. 1965. Studies on the pathogenesis of rheumatoid joint inflammation. I. The “RA cell” and a working hypothesis. Ann. Intern. Med. 62: 271-280. COIMBRA, A. & A. LOPEZ-VAS. 1967. Acid phosphatase positive cytoplasmic bodies in leucocytes of rheumatoid synovial fluid. Arthritis Rheumat. 10: 337-342. MARX,J. L. 1973. Slow viruses: role in persistent disease. Science 180: 13511354.
The earliest stages of rheumatoid arthritis ( R A ) have received little study, even though this may be the best time to search for mechanisms that initiate joint inflammation. Kulka et a / . in 1955 were impressed by the prominence of microvascular injury in their two earliest cases seen with suspected RA. We also described synovial vascular changes, including erythrocyte extravasation, thromboses, and neutrophil infiltration of venule walls, in our first report of synovitis of less than 1 month duration.' Both studies 1. showed some lining cell proliferation, even in the earliest cases, but a striking paucity of plasma cells. Synovial fluid leukocyte counts were often lower than in chronic RA. Mononuclear cells predominated in synovial effusions in many patients in these series and also in the report of Gatter and Richmond." This report describes a detailed study of the synovial membrane and the synovial fluid cells in eight patients seen during the first 6 weeks of arthritis, who now can be classified as definite or classic RA by the ARA criteria.' These patients are the only ones yet identifiable as RA out of a total of 104 patients seen since 1967 with arthritis of such recent onset. MATERIALS AND METHODS Eight patients initially without definite diagnoses but who now have evolved into classic RA are reported. TABLE1 lists their disease duration at time of synovial biopsy, initial rheumatoid factor titers, synovial fluid findings, and subsequent course. Each patient underwent a needle biopsy of the synovium with the ParkerPearson needle.5 Specimens were immediately placed into Bouin's solution for light microscopy and into half-strength Karnovsky's paraformaldehyde-glutaraldehyde fixative for electron microscopy. The synovial fluids were centrifuged, and the pellets were immersed in the same fixative. After 2 min of fixation, the synovial specimens and the exudate pellets were diced into 1 X l-mm pieces and replaced for a total of 4 hr, at room temperature, in the fixative solution. Specimens were washed in 0.1 M sodium cacodylate buffer at 4" C overnight. They were then postfixed in cold Palade's osmium-veronal for 2 hr, dehydrated in alcohol, placed in a 50% propylene oxide and EponB mixture for 2 hr, and then embedded in Epon 812. Thick sections ( 1 pm) were cut on an LKB-2
39
%
4
38
E
tl
z
P 0
41
Schumacher: Synovial Membrane Alterations
ultramicrotome with a glass knife, stained with toluidine blue, and examined for orientation. Thin sections were cut with a diamond knife and stained with uranyl acetate and lead citrate and then examined on an RCA EMU-3H electron microscope under a 50-kV beam. RESULTS Findings on light microscopy of the synovial biopsies are summarized in TABLE2. Note that some mild lining cell proliferation was apparent in all biopsies. At least occasional diffuse or perivascular lymphocytes were also seen in all specimens, and polymorphonuclear leukocytes (PMN) infiltrated the superficial synovium in six cases. Plasma cells were found in only patient 7, who had arthritis of 1-month duration. FIGUREl a shows a typical mildly
TABLE2 FINDINGS ON SYNOVIAL MEMBRANEBY LIGHTMICROSCOPY Patient 1 2 3 4 5 6 7 8
SLC
Layers 2-4 2-5 3-4 3-4 3-4 1-3 2-6 2-5
PMN 'r
Lymphocytes
2+
2+
+ + + 2+ + 2+ +
2+ 2+
2+ 3+
+
Plasma Cells
+
Vascular Obliteration
+ + + 3+ +
* SLC, synovial lining cells.
t PMN, polymorphonuclear leukocytes.
$
+, Definite but small numbers; 2 + ,
intermediate; 3 + , maximum.
involved area from patient 8 with P M N scattered throughout the center of the villus. FIGURE l b from patient 4 illustrates the type of vascular obliteration often seen. Vessels appeared to be plugged by inflammatory cells and also by organized thrombi. In FIGURE 2, we see more intense inflammation, including PMN and lymphocytes with narrow vessel lumens. Lumen narrowing possibly occurs merely because of temporary slower perfusion of this area. In this patient (no. 7 ) , other areas exhibited actual infiltration of vessel walls with PMN. This finding was the only suggestion of true vasculitis and is illustrated elsewhere (Reference 2, Figure 4). All changes were focal; some areas in several biopsies appeared perfectiy normal. Vascular congestion and edema were often noted. The synovial membrane electron microscopic findings are given in TABLE 3. There was no definite pattern for the type of synovial lining cell (SLC) that seemed to predominate. Type-A (phagocytic) and intermediate cells when present possessed long cytoplasmic processes and often contained finely granular
42
Annals New York Academy of Sciences
FIGUREla. Synovial villus from patient 8. There is dramatic vascular congestion, 2-3 layers of synovial lining cells (SLC), and a few scattered lymphocytes and polymorphonuclear leukocytes (arrow). Light microscopy; hematoxylin eosin stain. x 360.
material in their vacuoles. No virus-like particles were seen in SLCs (FIGURE 3). Occasional plasma cells were found by electron microscopy in two patients. Interstitial necrotic debris was common. I n all except one patient, vascular alterations were prominent. Gaps were noted between endothelial cells, and in these areas in particular, the endothelial cytoplasm displayed many flaps and processes that when cut on cross section might be superficially confused with 4 ) . Alterations involved both the superficial fenestrated viral particles (FIGURE venules and deeper thick-walled venules. Arteriolar disease was not observed. Vascular endothelium was frequently prominent and often showed rodlike Weibel-Palade bodies, as are also seen in normal vessels.x In the past, these bodies had been thought to be possible microorganisms.3fi There were also many filaments, occasional lipids, dense bodies, free ribosomes, cisternae of rough endoplasmic reticulum, pinocytic vesicles, mitochondria, multivesicular bodies, and normal appearing nuclei. Some endothelial cells had lost the normal cytoplasmic density and appeared to be degenerating. Four patients had some vessel lumens obliterated with platelet thrombi (FIGURE 5 ) . A single clump of paramyxovirus-like structures as seen in systemic lupus erythematosus was found in the vascular endothelium of patient 7.
Schumacher : Synovial Membrane Alterations
43
FIGUREI b. Superficial synovial vessel lumens plugged by inflammatory cells (arrows) from patient 4; light microscopy; hematoxylin and eosin stain. ~ 3 6 0 .
Endothelial basement membranes were often multilaminated. Pericytes and connective tissue components were frequently widely separated by low-densityareas, which suggested edema. There were extravasated erythrocytes and perivascular lymphocytes and PMN. PMN both perivascularly and within vascular lumens were degranulated, as previously described.!' I n four patients, particles were present that deserve consideration as to their possible viral nature. FIGURE 6 illustrates particles that appear to be budding from the vascular endothelium of patient 2. The particles in this patient are about 100-140 nm in diameter. FIGURE 7 from patient 6 also suggests a budding form: here, the smaller 90-nm budding structure seems to protrude into a vacuole. Each particle has a dense core surrounded by lucent areas and unit membranes and thus might be considered a C-type particle.lf1Il 1 Patients 6 and 8 also exhibited 80-140-nm particles in the perinuclear space without any definite budding. They did not have distinct central cores and 8 ) . Pacould also have been invaginations of endoplasmic reticulum (FIGURE tient 5 had 80-140-nm structures present in the vascular lumen near the endothelium, with a few central ribosome-like granules that are certainly not 9a). diagnostic but are identical in appearance to the arenaviruses * ? (FIGURE Many generally round particles with similar granules were also observed adjacent 9b). to a pericyte in patient 2 (FIGURE
C
+
+
PC t
+
2+
+
Surface Fibrin
+ ++ +
Viruslike Particles
t PC, plasma cells. I See footnote to TABLE 2 for definition of symbols.
* SLC, synovial lining cells; A, B, and C refer to types of lining cells.‘
7 8
5 6
4
2+
+ f
B
+ 2+ + 2+ 3+ 2+ 2+ + + + ++ 2 ++ 2+
+ + 3+
1
2 3
A
Patient
SLC 8
2+
+ 2+ ++ + 2+
Endothelial Gaps
+ + +
+ + ++ +
Dense Deposits in Vessel Walls Necrosis
TABLE3 ELECTRONMICROSCOPIC FINDINGS IN SYNOVIAL MEMBRANES
+++
+
Vascular Obliteration
+ +
Fibrin in Vessel Wall
s
2
2. 0
v,
?n
Y
Ea
2
P?
.c
5
E”
J
B
Schumacher : Synovial Membrane Alterations
45
FIGURE 2. Intensely inflamed synovium in patient 7. There are polymorphonuclear leukocytes, lymphocytes, and, in this case only, occasional plasma cells. Vessel lumens are narrowed. Light microscopy; hematoxylin and eosin stain. x 360.
TABLE 4 ELECTRONMICROSCOPY OF SYNOVIAL FLUIDCELLS Phagocytic Large Mononuclear Cells
Synthetic LM
Phagocytized Fine Granular Patient Cells and Debris Material I
2 3 4
5 6 I 8
+ + 2+
+ 31-
+ 2+ 2+ 2 f
%:
PMN Gray Globular Fine Granular Inclusions Inclusions
++ + + + 2t
* See footnote to TABLE 2 for definition of symbols.
+++ ++
+1-
46
Annals New York Academy of Sciences
FIGURE 3. Synovial lining cells (SLC) and superficial capillary of patient 5 with synovitis of 19 days’ duration. The most superficial SLCs and some of the deeper ones have prominent cytoplasmic processes, vacuoles, and dense bodies. There are varying amounts of rough endoplasmic reticulum (RER). These cells are type A (A) or intermediate (I). One type-B cell ( B ) with profuse RER is seen. V, small vessel without any identifiable pathologic change. x 6400.
Schumacher : Synovial Membrane Alterations
47
FIGURE 4. Synovial venule from patient 8 with gap (arrow) between two endothelial cells ( E ) . Many endothelial flaps and processes occur at the site of the endothelial junction opening. Note the loss of normal cytoplasmic density of the lower endothelial cell. This is probably a sign of cell degeneration. RBC, erythrocyte in the lumen; P, pericyte; BM, multilaminated basement membrane; electron micrograph. x 1 1,000.
48
Annals New York Academy of Sciences
FIGURE 5 . Platelets (PL) obliterate the lumen of a superficial synovial vessel from patient 6. G, gap between endothelial cells ( E ) ; BM, basement membrane; RBC, erythrocytes extravasated and lying in the vessel wall. The electron density of hemoglobin is not easily distinguishable from small dense deposits (D) of what may be other proteins in the vessel wall. P, pericytes; electron micrograph. x 12,800.
Schumacher: Synovial Membrane Alterations
49
FIGURE 6a. Virus-like particles (arrows) budding from the synovial vascular endothelium ( E ) of patient 2. Note that in this Figure, the endothelium is thin walled with a fenestration ( F ) . BM, basement membrane. Other particles free in the vessel lumen can be cell debris. Electron micrograph. x 48,400.
50
Annals New York Academy of Sciences
FIGURE 6b. See legend to FIGURE 6a. X 88,000.
In both patients 2 and 6, round particles with some central density were seen in vessel walls surrounded by electron-dense deposits (FIGURES 10a & b ) . Other cell debris that did not especially resemble viruses was also seen between endothelium and pericytes. Patients 4 and 6 also had occasional small dense deposits, without any identified particles between endothelium and pericytes. One other structure that merits consideration as possibly being virus related is shown in FIGURE 11. Here, tubules approximately 200 A in diameter lie in a parallel array in a nipple-like protuberance of a perivascular cell. This configuration is reminiscent of the nucleocapsids illustrated in budding virions of
Schumacher: Synovial Membrane Alterations
51
canine distemper virus l 1 and measles virus,I2 although the tubules are more densely packed in our patient. The known viral nucleocapsids have had smaller diameters of 100-150 A. Electron microscopy of the synovial fluid cells (TABLE 4) revealed no lymphocytes or other cells with paramyxovirus-like inclusions. PMN often contained gray globular bodies but very rarely any vacuoles of similar size with finely granular material (FIGUREI 2 a ) . F a r more evidence of phagocytosis was present in the large mononuclear (LM) cells; the LM cells of four patients contained large vacuoles filled with finely granular material, and five had phagocytized cell debris (FIGURE 12b). Four of the latter had ingested whole cells and were thus comparable to “Reiter’s” cells. No particles were especially suggestive of viruses or other organisms.
FIGURE 7. Ninety-nanometer particle (arrow) budding into an apparent vacuole in patient 6. PV, pinocytic vesicles; M, mitochondrion. The apparent particle in a small cisterna of rough endoplasmic reticulum (ER) may be an invagination of the ER; electron micrograph. x 32,000.
52
Annals New York Academy of Sciences
FIGURE 8. Perivascular cell with degenerating nucleus ( N ) showing loss of the nuclear envelope at the top. In this nucleus, as in many degenerating synovial nuclei in RA, gout, and other diseases, some of the chromatin appears as short tubules (arrows) that resemble the tubules in the tubuloreticular structures of systemic lupus erythematosus. Note also a variety of multivesicular bodies (MVB), unidentified round particles that may be cell debris, and particles in the perinuclear space (double arrows); electron micrograph. x 25,600.
Schumacher: Synovial Membrane Alterations
53
DISCUSSION In this review of an experience now expanded to include eight cases of rheumatoid arthritis seen within the first 6 weeks of disease, emphasis should be placed on the prominence of the microvascular changes. Microvascular injury is easily documented, but mechanisms of injury and the role of microvascular changes in perpetuation of arthritis are more difficult. All microvascular changes reported here (except for the budding virus-like particles) have also been observed in other acute inflammatory diseases, including transient undiagnosed synovitis.2 Microvascular obliteration, dense deposits, or other phenomena do not seem to permit any conclusions with respect to prognosis. However, it is of interest that those rheumatoid patients in this series who presented evidence of possible virus-like particles have had generally milder disease. All patients fit current criteria for RA and have as yet developed no excluding phenomena. Nevertheless, we should note that what is now known as RA may still include a wide variety of diseases with different initiating and perpetuating mechanisms, in addition to different prognoses. If microorganisms can initiate RA, one must suspect that many different organisms may be involved. Viral infection might cause only some “RA.” Because the group in this study was selected on the basis of relatively acute onset, it certainly is not representative of all RA. None of the patients who were latex negative when first examined have subsequently developed rheumatoid factor. We have studied the microvasculature to look for factors that might produce 10a and the initial injury. The virus-like particles are one possibility. FIGURES 10b from this series show virus-like particles embedded in electron-dense material in synovial vessel walls, which raises the intriguing possibility that viral antigen-antibody complexes produce the vascular injury. Such deposits are still far too few to do more than encourage continued search for similar areas in other patients. Virus particle-antiviral antibody complexes have been implicated in glomerular injury in lymphocytic choriomeningitis viral infection 34 and in the glomerulonephritis in murine leukemia.35 At present, we have no evidence of a direct cytopathic effect of any of the virus-like particles on synovial cells. This lack of evidence is in contrast to our work with hepatitis, where suspected viral particles were frequently associated with necrosis of involved ~ e l l s . 4 ~ Observation of occasional dense deposits in vessel walls without any viruslike particles also means that the presence of various other antigen-antibody complexes is still a possibility. Vessel lumens obliterated by platelets and packed with inflammatory cells did not invariably show any changes in the vessel wall. Thus, intraluminal processes might initiate some vascular disease. Other electron microscopic studies in RA have all been performed in wellestablished chronic disease and have rarely mentioned virus-like particle^.^^-^^ The tubular paramyxovirus-like structures so common in systemic lupus erythematosus are rare in RA.lX3143 Helder et d . 2 ’ found structures that suggest paramyxovirus nucleocapsids in only one of 25 RA synovial membranes. Gyorkey et d.lYreported such inclusions in each of four rheumatoid synovial membranes studied but did not describe their criteria for diagnosis or give ~ details about localization of the possible nucleocapsids. Neumark et ~ 1 . ’ examined 30 RA synovia from synovectomies and 17 noninflamed control membranes and found a variety of round particles 600 to 2500 A in diameter. Most were in the cytoplasm of vascular endothelium, macrophages, and other connec-
54
Annals New York Academy of Sciences
FIGURE 9a. Arenavirus-like particles (arrows) with unit membranes and ribosomelike granules. These particles could also not be distinguished from unusual cell processes in this endothelial cell adjacent to a gap ( G ) . 1,vascular lumen; patient 5; electron micrograph. x 49,600.
tive tissue cells. Convincing budding was not shown. Some clumps or isolated 160-1 80-Atubules in nuclei were described as paramyxovirus-like. More recent work by Neumark and FarkasZ3also suggests budding of virus-like particles from cell membranes, but cell membrane was not identifiable at the sites of suspected budding. Serre et aLZOreported but did not illustrate suspected viral nucleocapsids, 18 pm in diameter, in supernatants from tissue cultures of RA synovial membranes. have shown very slightly curved tubular 240-A diameter Helder et dZ4 structures not only in the cytoplasm of histiocytes in RA but also in other tissue
Schumacher : Synovial Membrane Alterations
55
reactions. These structures appeared identical to those found in patient 5 of our series and are not believed to be viral; they are, instead, considered similar to the membranous complexes seen, for example, in histiocytosis X.25 Helder et al. could find no particles like those reported by Neumark. Nuclear bodies in endothelial and perivascular cells have been described in rheumatoid synoviurn.l$ These authors recognized that none of them were specific for virus infection, but they did suggest a possible viral origin. We have not seen any nuclear material suggestive of viruses. Nucleoli13 might account for some of the images shown. Other tubules and filaments in crystallike arrays are similar to structures seen in the cytoplasm in many cells and are of unknown importance.:{h Most authors, including Neumark et al.,'2 have noted that it is difficult to distinguish viral particles from altered cell components, and this point deserves continued emphasis. Although electron micrographs can suggest the presence of viruses, culture will be required to prove viral invasion. Of course, even the culture of a virus does not establish its involvement in the pathologic process under consideration.
FIGURE 9b. Similar particles appearing to have arisen from a perivascular cell in patient 2; electron micrograph. X70,400.
56
Annals New York Academy of Sciences
FIGURE 10a. Dense deposit with virus-like particles beneath the endothelium (E) in patient 2. L, lumen; electron micrograph. x 70,400.
The rarity in RA of the paramyxovirus-like or tubuloreticular structures so typical of IupuslD seems to be a notable difference between these diseases. Recent cytochemical studies have suggested that the tubuloreticular structures may principally contain phospholipid and acid glycoprotein and thus most likely are not viral.51 Some tubular structures seen in our electron microscopic studies seem to be breakdown products of nuclear chromatin unrelated to viral 8 ) . Whether any of them have been considered virus-like infection (FIGURE tubules in the past is unknown. Definite identification of particles as viruses in electron microscopy thin sections is difficult, because a variety of cell components and still unidentified particles can superfically resemble virus particles.15 Most often, nonviral material resembles C-type viral particles, which have a dense core with a surrounding unit membrane. More intense osmium staining of the core or membrane, as 6 and 7 , tends to favor the presence of actual virus. illustrated in FIGURES Sources of possible confusion include transverse sections of collagen fibers, microvilli, and other cell processes (FIGURE4 ) , nuclear pores; annulate lamellae; vesicles in multivesicular bodies (FIGURE 8) ; secretion granules; pinocytic vesicles with dense inclusions, dense bodies, namely, lysosomes and platelet
Schumacher: Synovial Membrane Alterations
57
granules; mycoplasma elementary bodies; some hyperplastic endoplasmic re7 ) ; hyperplastic Golgi apparatus; and large nuclear granules. ticulum (FIGURE Dense bodies usually do not have the same wide space between the dense core and the membrane as viral C-type particles do. Membrane spikes observed on virus membranes at very high magnification also aid differentiation. Golgi apparatus, nuclear envelope, and pinocytic vesicles have all been described as contributing to the multivesicular bodies (MVB) .*';*li Their ubiquity suggests that they are not viral material, but it should not be considered proven that viruses cannot ever be components of MVB." The synovial fluid cell studies in our eight patients revealed no virus-like particles; this result is consistent with the hypothesis that if an infectious agent is involved, the initial insult occurs in the synovium. The dramatic amounts of cell debris present offer potential confusion with microorganisms, but
FIGURE lob. Particles, some of which could be cell debris in a loose appearing granular deposit ( D ) outside endothelium ( E ) in patient 6. Note also that the PMN in the lumen is almost completely degranulated; electron micrograph. x 11,200.
58
Annals New York Academy of Sciences
FIGURE 11. Nipple-like protuberance of a perivascular ribosome-rich cell of patient 2. Densely packed 200-A tubules lie in parallel array (arrows); electron micrograph. x 70,400.
Mitrovic et al.52 demonstrate how all these inclusions can be reproduced by phagocytosis of cell debris. As we have reported previously,Z PMNs in early RA synovial fluids do not seem to have any more than very rare electron microscopic evidence of phagosomes of the size described as cytoplasmic inclusions in “RA cells.’’53 The structures most consistent in size with the cytoplasmic inclusions of light and fluorescence microscopy appear to be the gray globular bodies described by Zucker-Franklin 4 3 that appear more likely to be lipid.5’ Granular material that resembles the complexes prepared by Zucker-Franklin is found much more commonly in the large mononuclear cells, some of which may be type-A SLC released into the joint space or monocyte-derived macrophages. Because of our hope than an infectious initiating agent might be demon-
Schumacher : Synovial Membrane Alterations
59
strable very early in disease, we have continued tissue culture studies since our 1971 report 26 and still can provide no evidence of viral replication. Slow viruses may be agents in such diseases as R A and are notoriously difficult to culture.55 Reasons for continued search for a virus or other infection that initiates R A include clinical observations of chronic polyarthritis that can follow presumed viral upper respiratory infection,27 experimental viral infection that produces chronic synovitis,23 the resistance of rheumatoid synovial cells to certain virus infections (reported but not yet confirmed) and the ability to transfer this resistance to rabbit synovial cells,zg and spontaneous chronic arthritis in chickens attributed to reovirus.:'O, 31 T h e intriguing association of ownership of pets, especially dogs, that antedates the onset of RA suggests
FIGURE 12a. Synovial fluid cells from patient 2. Lymphocytes at the bottom have prominent cytoplasm but no virus-like inclusions. Note the uropod (U).The three PMN above show structures like the gray globular (GG) bodies of Zucker-Franklin 4' that are very homogeneous and appear to be lipid. There are no comparably sized vacuoles with granular material. The small clearer areas at the arrows seem more like sites of degranulation than phagocytic vacuoles; electron micrograph. x 6400.
60
Annals New York Academy of Sciences
that an animal virus may cause the human disease. We have been studying a rheumatoid-like polyarthritis in dogs and have found tubuloreticular structures and crystalline arrays of 2 2 0 4 tubules but no particles comparable to those shown here in human RA.
FIGURE12b. Synovial fluid cells from patient 6. Large mononuclear cells contain vacuoles (V) with fine granular material. One cell has phagocytized a necrotic PMN farrows); electron micrograph. x 6400.
This work so far provides no clues to possible perpetuating mechanisms but does suggest some speculation. If a virus could arrive via the circulation to initiate the synovitis, some yet undetermined factor might determine whether viral antigen is sequestered in the joint to provide an ongoing antigenic stimulus. Rubella virus antigen, for example, has been found in chondrocytes of con-
Schumacher : Synovial Membrane Alterations
61
genitally infected rabbits.4* Synovial and cartilage cells can be infected by viruses in ~ i t r - 0 Organisms .~~ may well not be morphologically identifiable, even though they still provide antigen. In known infections, as with mycoplasma hyorhinis infection of swine,32 it is important to note that organisms may not be identifiable by electron microscopy, despite the fact that they can still be ~ shown that antigen in cultured from the joint. The work of Cooke et U Z . ~has rabbits sensitized to bovine serum albumin (BSA) and then given intraarticular BSA is sequestered in cartilage and other connective tissue. Some initial local inflammatory reaction seems to be necessary in a yet undefined way to allow the complexes to enter and be stored in the collagenous tissues.47 Cell debris, and therefore possibly viral antigen, is scattered throughout the synovial collagenous tissue in our electron micrographs. SUMMARY Eight patients have been studied during the first 6 weeks of rheumatoid synovitis. All of them exhibited microvascular injury, which was manifested by gaps between endothelial cells, vascular occlusion, erythrocyte extravasation, or endothelial cell injury. In four patients, a variety of virus-like particles were found associated with the endothelium or perivascular cells. In two cases, particles were seen in electron-dense deposits in vessel walls. Lymphocytes and P M N infiltrated the synovial membranes, but plasma cells were uncommon. Evidence of phagocytosis was prominent in synovial lining cells and other large mononuclear cells, but not in PMN. These observations are consistent with injury to synovium and, specifically, synovial vessels as an early stage in RA synovitis. The virus-like particles require further investigation, because nonviral cell components remain very difficult to distinguish in electron microscopy tissue sections. REFERENCES
J. P., D. BOCKING,M . W. ROPES& W. BAUER.1955. Early joints lesions 1. KULKA, of rheumatoid arthritis. Arch. Pathol. 59: 129-150. 2. SCHUMACHER, H. R. & R. C. KITRIWU.1974. Synovitis of recent onset. A clinicopathologic study during the first month of disease. Arthritis Rheumat. 15: 465-485. 3. GATTER,R. A. & J . D. RICHMOND. 1973. The predominance of synovial lymphocytes in early rheumatoid arthritis. Arthritis Rheumat. 1 6 544 (abs.). S. COBB,R. JACOX & R. A. JESSAR.1958. Revi4. ROPES,M. W., G. A. BENNETT, sion of diagnostic criteria for rheumatoid arthritis. Bull. Rheumat. Diseases 9: 175, 176. 5. SCHUMACHER, H. R. & J. P. KULKA.1974. Needle biopsy of the synovial membrane. Experience with the Parker-Pearson technic. N. Engl. J. Med. 286: 41 6-4 19. 6. KARNOVSKY, M. J. 1965. A formaldehyde-glutaraldehyde fixative of high osrnolality for use in electron microscopy. J. Cell Biol. 27: 441 (abs.). 7. BARLAND, P., A. B. NOVIKOFF & D. HAMERMAN. 1962. Electron miscroscopy of the human synovial membrane. J . Cell Biol. 14: 207-220. H. R. 1969. The microvasculature of the synovial membrane of 8. SCHUMACHER, the monkey. Ultrastructural studies. Arthritis Rheumat. 12: 387-404. H. R. & C. A. AGUDELO.1972. Intravascular degranulation of 9. SCHUMACHER, neutrophils: an important factor in inflammation? Science 175: 1139, 1140.
62
Annals New York Academy of Sciences
W. 1960. The detection and study of tumor viruses with the electron 10. BERNHARD,
microscope. Cancer Res. 20: 712-727. 1 1 . DALTON,A. J. 1972. Further analysis of the detailed structure of type B and C particles. J. Nat. Cancer Inst. 48: 1095-1097. 12. DALTON,A. J. 1973. The arenaviruses. In Ultrastructure of Animal Viruses and Bacteriophages. A. J. Dalton & F. Haguenau, Eds. : 323. Academic Press, Inc. New York, N.Y. D. W. 1966. The Cell. : 26. W. B. Saunders Company. Philadelphia, 13. FAWCETT, Pa. 1970. Nuclear bodies in rheumatoid synovium. T. & K. FARKAS. 14. NEUMARK, Ann. Rheumat. Diseases 29: 653-659. 15. HAGUENAU, F. 1973. “Virus-like’’ particles as obzerved with the electron microscope. I n Ultrastructure of Animal Viruses and Bacteriophages. A. J. Dalton & F. Haguenau, Eds. : 391-396. Academic Press, Inc. New York, N.Y. 16. MATIN, B. J. & S. S. SPICER.1973. Multivesicular bodies and related structures of the syncytiotrophoblast of human term placenta. Anat. Rec. 175: 15-34. W. & A. JASINSKI.1970. The formation of multivesicular bodies 17. KILKARSKI, from the nuclear envelope. J. Cell Biol. 45: 205-21 I . 18. SCHUMACHER, H. R. 1970. Ultrastructure of the synovial microvasculature in rheumatoid arthritis and systemic lupus erythematosus. I n Proceedings of the Electron Microscopy Society of America. C. I. Arceneaux, Ed. : 240. Claitor’s Publishing Division. New Orleans, La. K. W. MIN & P. GYORKEY.1972. A morphologic F., J. G. SINKOVICS. 19. GYORKEY, study on the occurrence and distribution of structures resembling viral nucleocapsids in collagen diseases. Amer. J. Med. 53: 148-158. 20. SERRE,H., J. MANDIN, C. VAGO,J. SANG& J . CLOT. 1972. Recherche et etude de virus dans la synoviale au cours de la polyarthrite chronique rhumatismale. Rev. Rhumat. 39: 595-599. 21. SCHUMACHER, H. R. 1973. Needle synovial biopsy: evidence for infectious agents including viruses as possible etiologies in arthritis. Excerpta Med. 299: 27 (abs.). T., I. HOLLOS& K. FARKAS. 197 I . Electron microscopical investiga22. NEUMARK, tion of virus-like particles in rheumatoid synovial membrane. Vnitini Lekaistri. 17: 625-630. 23. NEUMARK,T. & K. FARKAS.1973. Ultrastructural aspects of lymphoreticular cells in rheumatoid synovium. Ann. Rheumat. Diseases 32: 524-530. 24. HELDER,A. W., T. M. FELTKAMP-VROOM & R. L. F. NIENHUIS.1973. Electron and light microscopical observations and serological findings in rheumatoid arthritis. Ann. Rheumat. Diseases 32: 5 15-523. 25. BASSET,F., J. ESCAIG& A. LECROM.1972. A cytoplasmic membranous complex in histiocytosis X. Cancer 29: 1380-1386. 26. SCHUMACHER, H. R., S. L. KATZ& R. B. KUNDSIN.197 I . The continued search for viral and other infectious agents in the systemic rheumatic diseases: study of early lesions. In Atypical Virus Infections. C. L. Christian, Ed. : 18-23. The Arthritis Foundation. New York, N.Y. 27. HOLLANDER, J. L., E. M. BROWN,R. A. JESSAR,K. HUMMELER & W. HENLE. 1962. Studies on the relationship of virus infections to early or acute rheumatoid arthritis. Arch. Interamer. Rheumatol. 5: 137-157. 28. WEBB, R. W., R. BLUESTONE,L. S. GOLDBERG, S. D. DOUGLAS & C. PEARSON. 1974. Experimental viral arthritis induced with herpes simplex. Arthritis Rheumat. 1 6 241-250. 29. SMITH,C., D. HAMERMAN, R. JANIS& E. HABERMAN.1974. Virus resistance transferred from human rheumatoid cells to rabbit synovial cells. Ann. Rheumat. Diseases 33: 173-179. 30. GLASS,S. E., S. A. NAQI, C. F. HALL & E. M. KERR. 1973. Isolation and characterization of a virus associated with arthritis of chickens. Avian Diseases 17: 415-424.
Schumacher : Synovial Membrane Alterations
63
E. R., M. H. FRIEDMAN & N . 0. OLSON. 1972. Electron microscopic 31. WALKER, study of an avian reovirus that caused arthritis. J . Ultrastruct. Res. 41: 67-69. J. R. & R. F. Ross. 1969. Fine structure of synovial membrane in 32. DUNCAN, mycoplasma hyorhinis arthritis of swine. Amer. J. Pathol. 57: 171-181. H. R. 1970. Tubular paramyxovirus-like structures in synovial 33. SCHUMACHER, vascular endothelium. Ann. Rheumat. Diseases 29: 445-447. M. 9. A. & F. J . DIXON.1970. Pathogenesis of chronic disease 34. OLDSTONE, associated with persistent lymphocytic choriomeningitis viral infection. 11. Relationship of the anti-lymphocytic choriomeningitis immune response to tissue injury in chronic lymphocytic choriomeningitis disease. J. Exp. Med. 131: 1-19. 35. PASCAL,R. R., M. N . Koss & R. L. KASSEL. 1973. Glomerulonephritis associated with immune complex deposits and viral particles in spontaneous murine leukemia. An electron microscopic study with immunofluorescence. Lab. Invest. 29: 159-165. T. C. & D. G. RAYNS. 1966. Further electron microscopic studies of 36. HIGHTON, the human synovium. Proc. Univ. Otago Med. School 44: 37-39. & D. HAMERMAN. 1964. Fine structure of the 37. BARLAND,P., A. 9. NOVIKOFF rheumatoid synovial membrane with special reference to lysosomes. Amer. J. Pathol. 44: 853-866. 38. GHADIALLY, R. N. & S. ROY. 1967. Ultrastructure of synovial membrane in rheumatoid arthritis. Ann. Rheumat. Diseases 26: 426-443. 39. NORTON,W. L. & M. ZIFF. 1966. Electron microscopic observations on the rheumatoid synovial membrane. Arthritis Rheumat. 9: 589-609. 1971. UltraJ., A. VALLE,L. DIAZ-FLORES & J. M. ARAMBURU. 40. BORRACHERO, estructura de la sinovitis reumatoide. Rhumatologie 1: 277-285. H. J. C., H. M. LAIRD,N. G. WRIGHT & H. THOMPSON.1971. 41. CORNWELL, Ultrastructural features of canine distemper virus infection in a dog kidney cell line. J . Gen. Virol. 12: 281-292. 42. NAKAI,T., F. L. SHAND& A. F. HOWATSON.1969. Development of measles virus in vitro. Virology 38: 50-67. 43. ZUCKER-FRANKLIN, D. 1966. The phagosomes in rheumatoid synovial fluid leucocytes. A light, fluorescence and electron microscopic study. Arthritis Rheumat. 9: 24-36. 44. LONDON, W. T., D. A. FUCCILLO, 9. ANDERSON & J . L. SEVER.1970. Concentration of rubella virus antigen in chondrocytes of congenitally infected rabbits. Nature (London) 226: 172, 173. 45. CRAIG,C. P. & R. B. GREER. 1971. Virus infections in rabbit articular cells. Clin. Res. 19: 439. T. D., E. R. Hum, M. ZIFF & H. E. JASIN. 1972. The pathogenesis of 46. COOKE, chronic inflammation in experimental antigen-induced arthritis. 11. Preferential localization of antigen-antibody complexes to collagenous tissues. J. Exp. Med. 135: 323-338. 47. COOKE,T. D. & S. M. RICHER. 1974. Pathogenesis of chronic inflammation in experimental antigen-induced arthritis. 111. Influence of immune state and local inflammatory response on early localization of antigen to articular collagenous tissues. 1. Rheumat. 1: 143-152. H. R. & E. P. GALL.1974. Arthritis in acute hepatitis and chronic 48. SCHUMACHER, active hepatitis. Pathology of the synovial membrane with evidence for the presence of Australia antigen in synovial membranes. Amer. J. Med. 57: 655-664. N. L., N . DITCHEK, J. POILEY& I. N. KIEM. 1974. Pets and 49. GOTTLIEB, rheumatoid arthritis. Arthritis Rheumat. 17: 229-234. C. D., H. A. ALLEN,R. E. HALLIWELL & H. R. SCHUMACHER. 1974. SO. NEWTON, Clinico-pathologic conference. J. Amer. Vet. Med. Ass. 165: 459-464. 51. SCHAFF, Z., D. BARRY & P. M. GRIMLEY.1973. Cytochemistry of tubuloreticular structures in lymphocytes from patients with systemic lupus erythematosus
64
52.
53. 54.
55.
Annals New York Academy of Sciences and in cultured human lymphoid cells. Comparison to a pararnyxovirus. Lab. Invest. 29: 577-586. MITROVIC,D., J. BARBOT,M. F. KAHN,D. Hroco & S. DE SBZE. 1969. Reproduction experimentale des inclusions gtantes dans les macrophages pulmonaires du lapin. Pathol. Biol. Sern. Hop. 17: 53-58. HOLLANDER, J . L., D. J. MCCARTY,G. ASTORGA& E. CASTRO-MURILLO. 1965. Studies on the pathogenesis of rheumatoid joint inflammation. I. The “RA cell” and a working hypothesis. Ann. Intern. Med. 62: 271-280. COIMBRA, A. & A. LOPEZ-VAS. 1967. Acid phosphatase positive cytoplasmic bodies in leucocytes of rheumatoid synovial fluid. Arthritis Rheumat. 10: 337-342. MARX,J. L. 1973. Slow viruses: role in persistent disease. Science 180: 13511354.
