J Clin Pathol 1983;36:734-741

Heterogeneity of HLA-DR+ cells in normal humankidney. Immunohistological and cytochemicalcharacterisation of discrete cell populationsMARTIN J RAFTERY, LEONARD W POULTER,* GEORGE JANOSSY,* PAUL SWENY,OSWALD N FERNANDO, JOHN F MOORHEAD

From the Department ofNephrology and Transplantation, Royal Free Hospital, London NW3 2QG and the*Department ofImmunology, Royal Free Hospital School of Medicine, London NW3 2QG

SUMMARY Biopsies of normal kidneys taken at time of transplantation were studied using a

variety of immunofluorescent and cytochemical techniques. A heterogeneous population ofHLA-DR+ cells was found, mainly confined to the intertubular interstitium. The majority ofthese cells (80%) were positive when stained with a rabbit anti-factor VIII antiserum suggestingthat they were endothelial cells. A minority however (20%) were factor VIII- but were positivelystained with FMC17, a monoclonal antibody (McAb) directed against human monocyte/macrophage antigens. Positive staining of this subpopulation was also noted with RFD1, a McAbwhich reacts with an antigen on human interdigitating cells (ID cells). Cytochemical reactionsrevealed that these cells contain adenosine triphosphatase (ATPase) and acid phosphatase(ACP) and thus do not conform to the phenotype of tissue histiocytes. The phenotype of thislatter population is identical with that of the ID cells found in tonsil, thymus and spleen and it issuggested that they play a major role in initiating the process of renal allograft rejection.

Acute graft rejection is responsible for over 75% ofrenal allograft failures.' 2 The graft rejection processis the result of host immunological mechanismsresponding to alloantigens expressed by the cells ofthe donor kidney. Despite the fact that a majority ofkidney grafts are now performed with a closeHLA-ABC match, many grafts are still lost due torejection. A close HLA-DR match between donorand recipient has been shown to improvesignificantly the one-year graft survival.34 Thisobservation suggests therefore that the HLA-DRantigens may also play a significant part in initiatingand driving the rejection process. In rats relativelylarge numbers of Ia+ (HLA-DR equivalent) cellshave been identified in the interstitium of the nor-mal rat kidney.5 These cells are stellate in appear-ance and may form part of a population of Ia+ cellswidely distributed in all tissues in the rat, except thebrain.6 In terms of their morphology, surface anti-gens and cytochemical activity these cells are indis-tinguishable from the "dendritic" cells identified inthe mouse, which have been shown to have the func-tional capacity of antigen presentation.78 Batchelor

Accepted for publication 12 January 1983

and his coworkers have tested the hypothesis thatthese cells play a role in the immunological responseto rat renal allografts.9 They have shown that rejec-tion of a long surviving immunologically enhancedkidney allograft could be provoked by infusion of anenriched suspension of donor strain dendritic cellsinto the recipient indicating that Ia+ dendritic cells(but not Ia+ B cells) have a major role in the activa-tion of the graft rejection process.Homogenates of normal human kidney have been

shown to contain relatively large amounts ofHLA-DR antigens.'" A population of cells in theintertubular interstitium express HLA-DR antigenson their surface membranes." 12 However,HLA-DR antigens are expressed on a wide range ofcell types including both lymphoid and non-lymphoid populations.'2"'4 A full understanding ofthe relevance of HLA-DR+ populations in renalallograft rejection demands an analysis of the celltypes involved and their microenvironmental rela-tions both in normal kidney and in rejecting allo-grafts. Discrete characterisation of unique cell typeswithin the tissue sections has recently been shownpossible by the use of combinations of immunohis-tological and cytochemical techniques in frozen sec-

734

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Heterogeneity ofHLA-DR+ cells in normal human kidney

tions'3 and this paper describes such an analysis ofnormal human kidney.

Material and methods

SPECIMEN PREPARATIONNeedle biopsies were taken from normal humankidneys at the time of transplantation using thestandard Trucut technique. The biopsies weremounted on cork, embedded in O.C.T. (AmesCorp) and frozen in isopentane cooled in a bath ofliquid nitrogen. Six micron sections were cut in acryostat maintained at -25°C, air dried for 2 h andfixed for 5 min in chloroform:acetone (1:1). Theywere freeze-dried for 2 h, individually wrapped andmaintained at -20°C until required for staining.These techniques have been described in detail.'3

ANTIBODIES AND CYTOCHEMISTRYThe types of antibodies used, their sources and stain-ing characteristics are described in Table 1. Doubleimmunofluorescent staining was performed usingsecond layer antisera conjugated with fluoresceinisothiocyanate (FITC) and tetraethyl rhodamineisothiocyanate (TRITC). Second layer antisera weremade in goats (G), sheep (S) or swine (Sw) againstthe relevant Ig and were either eluted fromimmunoadsorbant columns or were IgG fractions ofhyperimmune antisera. They were then conjugatedto the appropriate fluorochrome (for example, FITCor TRITC) and were standardised on tissue sections(for example, tonsil) using first layer reagents ofknown activity and staining patterns. In the majorityof experiments described below double labelling wasperformed using chicken anti Ta complexed to sheepantichicken FITC and the appropriate mouse mono-clonal antibody complexed to goat antimouse IgTRITC.

Acid phosphatase activity (ACP) and adenosinetriphosphatase activity (ATPase) were demons-trated using established techniques."I In some casescytochemical staining was performed on tissue sec-tions previously stained by immunofluorescent tech-niques (IF). After IF staining the slides were washedin phosphate-buffered saline (PBS) for 10 min, thenin distilled water for 10 min. They were then fixed inice-cold calcium formalin for 3-5 min, again washedin distilled water and then incubated in the approp-riate cytochemical incubation medium. The sectionswere examined on a standard Zeiss microscope withan epifluorescence condenser containing selectivefilters for FITC and TRITC.

EXPERIMENTAL CONTROLSFrozeni sections of tonsil, thymus, lymph node andsynovial membrane were used as control tissues to

validate the staining characteristics of the variousantibodies used. Sections of kidney were incubatedwith irrelevant antibody and with 2nd and 3rd layerantisera alone to control for non-specific staining.

Results

LOCALISATION OF HLA-DR+ CELLSIndirect immunofluorescence techniques usingeither heterologous anti-HLA-DR antisera ormonoclonal anti-HLA-DR antibody revealed a con-sistent pattern of positive cells in all sections studied.The HLA-DR+ cell population were found almostexclusively in the intertubular interstitium (Fig. 1).This localisation occurred in sections of both cortexand medulla. Glomerular capillary loops andmesangial cells were also HLA-DR+ (Fig. 2d).There was no membrane staining of tubular epithel-ial cells, although there was a very faint backgroundcytoplasmic fluorescence.

MORPHOLOGY OF HLA-DR+ CELLSThe morphology of the HLA-DR+ cells appearedextremely heterogeneous, even allowing for the factthat tissue sectioning may cut cells at different ang-les. Three fundamentally different cell shapes couldbe identified (Table 2). The majority of cells wereeither round or crescent-shaped, being elongatedand curved in close apposition to kidney tubules andwere relatively small in size. A minority were largerand stellate in appearance and some of the lattercells had processes which extended between kidneytubules. No obvious relation was found betweenanatomical localisation and morphology. Cells of allmorphological types were found in all areas of thekidney, although there was a preponderance of cellsof an interdigitating morphology in the cortex.Examples can be seen in Figs. 1, 2b and 2d, 4a and4c, and 5a.

IMMUNOLOGICAL CHARACTERISATIONTo characterise the HLA-DR+ cells, combinationsof immunohistological and cytochemical reactionswere performed on the same or serial sections.

Reagents detecting lymphocyte surface markers(eg BAl and UCHT1) failed to react with theHLA-DR+ cells. Conversely, W6/32 a mousemonoclonal antibody directed against HLA-ABCantigens "stained" virtually all the HLA-DR+ cells.When both reagents (HLA-DR and HLA-ABC)were tested in combination on the same sections, analmost complete concordance of reaction patternwas seen in all morphological areas of the kidney(Fig. 2a and 2d). Although all HLA-ABC+ cellswere HLA-DR+, occasional individual cells wereHLA-DR+, HLA-ABC-.

735

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Raftery, Poulter, Janossy, Sweny, Fernando, Moorhead

Fig. 1 Kidney cortex stained with ananti-HLA-DR McAb and goatanti-mouse conjugated to peroxidase assecond layer. A light haematoxylincounterstain has also been applied. Theglomerular capillary loops areHLA-DR' as are the linear peritubularcapillaries. Interspersed in the interstitiumare occasional HLA-DR' cells ofmorecomplex morphology (examplesarrowed).

Fig. 2 Normal kidney stained with a combination of W6132 a McAb against human HLA ABC antigens, (Fig 2a and c)and chicken anti HLA-DR (Fig 2b and d). Second layer reagents were goat-anti mouse TRITC and sheep ant chicken FITCand they were photographed using filters selective for TRITC and FITC. Fig. 2a shows an area ofmedulla with HLA-ABC+cells arranged longitudinally between tubules and Fig. 2b shows the same field labelled with ant HLA-DR. Fig. 2c shows aglomerulus stained with chicken anti HLA-DR. There is almost 100% concordance between HLA-ABC+ and HLA-DR+cells.

736

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Heterogeneity of HLA-DR+ cells in normal human kidney

Fig. 3 Normal kidney double stained with anti HLA-DR McAb and rabbit ant factor VIII RAg. Picture 3a (taken througha fluorescein sensitive filter) shows part ofa glomerulus and some proximal tubules. The glomerular capillary loops and theintertubular cells are HLA-DR+. Picture 3b (taken through a rhodamine sensitive fiter) shows the same field in which thereis confpuent positive staining of the glomerular endothelium and some, though not all of the intertubular HLA-DR+structures are stained.Table 1 Antibodies used in this study

Designaton Source Specificity Reference

Chicken heterologous Chicken heterologous antiserum B cells Janossy G et all'antiserum to purified human p28,33 antigens Some activated T cells

Interdigitating cellsSome macrophages

OKla Mouse monoclonal antibody As above Reinherz EL et al'5against human p28,33 antigens

YE2/36 Mouse monoclonal antibody As above Bricknell PM et al'7against human p28,44 antigens

FMC-17 Mouse monoclonal antibody Circulating monocytes Seymour GJ et all'against human macrophage and Cells of macrophage/monocytemonocyte antigen lineage

Rabbit anti-factor VIII Rabbit heterologous antiserum Endothelial cells Dako Labagainst human factor VHIRAg

W6/32 Mouse monoclonal antibody All leucocytes Barnstable CJ et a"l'against human HLA-ABC antigens Most epithelial cells

RFD1 Mouse monoclonal against ID cells in lymph node, Rawlings Bofill et alhuman ID cells tonsil and thymus. 20% (in preparation)

of B cells, 3-4% ofimmature myeloblasts

BAI Mouse monoclonal against B cells in peripheral blood Abrahamson CS et al2'human B cells lymph nodes and tonsil B

cell precursors in bone marrowUCHTI Mouse monoclonal against T cells in peripheral blood Callard RE et a!2'

human T cells (pan T) lymph node and thymus

737

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Raftery, Poulter, Janossy, Sweny, Fernando, Moorhead

Table 2 HLA-DR+ cells in normal kidney

Morphology HLA-ABC Factor VIII RFDI FMC17 ACP ATPase

Crescent-shaped, + + .- - - +round, or ovalStellate + - + + - +

Combined reactions using anti-HLA-DR andanti-factor VIII antisera revealed that the HLA-DR" cells were heterogeneous not only in terms ofmorphology (see above) but also in terms of surfacemarkers. Anti-factor VIII antiserum demonstrated anetwork of positive cells interspersed between thetubules corresponding in location to the capillaryendothelium (Fig. 3a and 3b). A proportion ofHLA-DR+ cells (approx 20%) were factor VIII-and these were predominantly the large cells in theintertubular interstitium. The close anatomical

association of all cells in the intertubular interstitiumand the fact that the factor VIII staining was cytop-lasmic made it difficult to measure the exact propor-tions of HLA-DR+ cells that were either factorVIII+ or factor VIII-.

Clarification of the picture emerged whenHLA-DR+ cells were stained with FMC17, a McAbdetecting cells of the monocyte/macrophage series.While the majority of DR' cells were FMC17-, aproportion (approx 20%) were FMC17+ (Fig. 4band d). Combined staining using anti-HLA-DR

4;t

Fig. 4 Sections ofnormal kidney stained with anti-HLA-DR antiserum, FMC-17 McAb and Mg++ dependent A TPase.Picture 4a (taken through a fluorescein sensitive filter) shows a selection ofHLA-DR+ cells. Picture 4b (taken through arhodamine sensitve filter) shows the same field in which two ofthe HLA-DR+ cells (arrowed) are unequivocally FMC- 17+.Picture 4c shows several HLA-DR+ cells and a large blood vessel (labelled*). Picture 4d shows the same field which hasbeen subsequently labelled with A TPase. The vessel wall is seen to be strongly ATPase+ andpractically all ofthe HLA -DR +cells are also ATPase+.

738

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Heterogeneity ofHLA-DR' cells in normal human kidney

Fig. 5 Normal kidney stained withanti HLA-DR antiserum andRFD]. Picture 5a taken throughan FITC sensiiive filter showsnumerous DR+ cells some of which(labelled*) are clearly endothelial.Picture 5b (taken through aTRITC sensitive filter) shows thesame field labelled with RFDI inwhich the endothelial cells (*) areclearly negative and in whichseven cells with an interdigitatingmorphology (arrowed) are clearlyRFDI+.

antiserum and RFD1 (a McAb detecting human IDcells) confirmed that approximately 20% of theDR' cells were RFDI+ and these were clearly of astellate or interdigitating morphology (Fig. 5a and5b). It was also clear that endothelial cells wereRFDl-.

COMBINED CYTOCHEMICAL REACTIONSSets of sections from different biopsies of normalkidney were "stained" with HLA-DR antiserumand then placed in incubation medium to demons-trate enzyme activity. Using this technique it was

found that the HLA-DR+ cells in the interstitiumfailed to express any ACP activity irrespective oftheir localisation, morphology or other immunohis-tological characteristics. Occasional isolated kidneytubules were ACP+. In contrast, combined reactionsto detect ATPase activity demonstrated that all theHLA-DR+ cells expressed discrete membrane activ-ity for this enzyme (Fig. 4c-d).

Discussion

The results described here reveal that normal kidney

739

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Raftery, Poulter, Janossy, Sweny, Fernando, Moorhead

contains a significant population of cells expressingHLA-DR antigens and furthermore that this popu-lation is heterogeneous. The presence of HLA-DRantigens in homogenates of kidney has been previ-ously reported although this study failed to localisethe HLA-DR expressing cells.'0 Koyama et aldemonstrated HLA-DR+ cells in the intertubularinterstitium and in the glomeruli which were thoughtto be endothelial cells," while a further studydemonstrated HLA-DR+, HLA-ABC+ cells in theinterstitium, a proportion of which were thought tobe dendritic cells analogous to the dendritic cellsdescribed by Steinman and his co-workers in themouse.7 12 However, this subpopulation was notanalysed for additional immunohistological andcytochemical markers which are characteristic ofdendritic cells in other locations.The morphology and phenotype of the HLA-

DR' cells which are factor VIII+, RFDl-, FMC17-,ATPase+ described above strongly supports thesuggestion of Koyama et al that the majority of'HLA-DR+ cells are indeed endothelial. However,approximately 20% of the HLA-DR+ populationare factor VIII- but FMC17+, RFD1+. Theseimmunocytochemical characteristics, together withthe stellate morphological appearance of these cells,offers strong evidence that they are indeed dendriticor interdigitating cells. Their cytochemical reactions(that is, ATPase+, ACP-) are at variance with those-of HLA-DR+ activated macrophages and providesfurther'evidence that this is a unique cell population.This cell type is found in the paracortical area of the'lymph node and the corticomedullary junction ofthe thymus.'62022 These cells also accumulate in thesynovia of patients suffering from active rheumatoidarthritis and in the skin during delayed typehypersensitivity reactions.2324 It has been suggestedthat such cells are analogous to the dendritic cells of,mice that are known to possess antigen presentingcapacity.82225

While there is good evidence for the implicationof dendritic cells of donor origin in allograft rejec-tion in rodents, this has not been established in thehuman.9 There is indirect evidence from retrospec-tive reviews of allograft survival, that DR antigenshave a crucial role in the ultimate acceptance or

rejection of the allograft.3426 However, this begs thequestion as to which population of DR' cells areimplicated and whether all of the DR' cells havesimilar capacities for antigen presentation. Althoughthe HLA-DR antigens on the capillary endotheliummay act as target for host derived rejection proces-

ses, it is tempting to speculate that the possible anti-gen presenting capacity of the above describedHLA-DR+, FMC17+, RFDl+ ATPase interdigitat-ing cells of the donor kidney may trigger the early

episodes of acute graft rejection. Weight is added tothis hypothesis by the focal nature of the interstitialcellular infiltrates that accompany the acute rejec-tion process.

Confirmation of the involvement of HLA-DR+ID cells in human renal allograft rejection is stillawaited but it would seem that the above descriptionof their phenotype and localisation, should provide afirm basis for further elucidation of their role in thein situ rejection process.

Rf*emm

'Burkeson RL, Marbarger PD, Schroder ET, Brennan AM. Thehazards of renal transplantation. Kidney Int 1982;21:293.

Brynger H, Brunner FP, Chantler C. Combined report on regulardialysis and transplantation in Europe X. Proc Eur Dial'Transplant Assoc 1980;172-84.

3 Morris PJ, Bishop M, Fellows G, et al. Results from a new renalltransplantation unit. Lancet 1978;ii: 1353-6.

4Persijn GG, Van Leeuwem A, Parievliet J, et al. Two majorfactors influencing kidney graft survival: HLA-DR matchingand blood transfusion. Transplant Proc 1981;13:150-4.

s Hart DNJ, Fabre JW. Major histocompatibility complex antigensin rat kidney ureter and bladder. Transplantation1981;31:318-25.

6Hart DNJ, Fabre JW. Demonstration and characterization of Iapositive dendritic cells in the interstitial connective tissues ofrat heart and other tissues but not brain. J Exp Med1981;154:347-61.

Steinman RM, Cohn ZA. Identification of a novel cell type inperipheral lymphoid organs of mice I. J Exp Med1973;137:1142-62.

Steinman RM, Cohn ZA. Identification of a novel cell type inperipheral lymphoid organs of mice II. Functional propertiesin vitro. J Exp Med 1974;139:380-92.

9 Lechler RI, Batchelor JR. Restoration of immunogenicity to pas-senger cell depleted kidney aliografts by the addition of donorstrain dendritic cells. J Exp Med 1982;155:31-41.

'° Williams KA, Hart DNJ, Fabre JW, Morris PJ. Distribution andquantitation of HLA-ABC and DR antigens on human kidneyand other tissues. Transplantaton 1980;29:274-9.

Koyama K, Fukunishi T, Burcos M, Tanigaki M, Pressman D.Human Ia-like antigens in non-lymphoid organs. Immunology1979;38:333-41.

12 Hart DNJ, Fuggle SF, Williams KA, Fabre JW, Ting A, MorrisPJ. Localization of HLA-ABC and DR antigens in humankidney. Transplantaion 1981;31:428-33.

'3 Poulter LW, Chilosi M, Seymour GJ, Hobbs S, Janossy G.Immunofluorescence membrane staining and cytochemistryapplied in combination for analysing cell interactions in situ.In: Polak J, van Noorden S, eds. Immunocytochemistry; practi-cal applications in pathology and biology. London: Wright,1983:233-48.

Natali PG, Martino CD, Quaranta V, et al Expression of Ia-likeantigens in normal human non-lymphoid tissues. Transplanta-tion 1981;31:75-8.

5 Reinherz EL, Kung PC, Pesando JM, Ritz J, Goldstein G,Schlossman SF. Ia determinants on human T cell subsetsdefined by monoclonal antibody: activation stimuli requiredfor expression. J Exp Med 1979;15O:1472-82.

16 Janossy G, Thomas JA, Goldstein G, Bollum FJ. The humanthymic microenvironment. In: Microenvironments inhaemopoietic and lymphoid differentiation. Ciba FoundationSymposium 84. London: Pitman Medical, 1981:193-213.

Bricknell PM, McConnell I, Milstein C, Wright B. A monoclonal

740

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from

Heterogeneity ofHLA-DR+ cells in normal human kidney

andtbody to the HL4A-DR product recognises a polymorphic ladeterminant in mice. Immunology 1981;43:493-501.

"Seymour GJ, Poulter LW, BofiHl M, et al. The reactivity of amonoclonal antibody against cells of the monocyte/macrophage series in normal and inflamed human tissues. JRetic Soc 1983; (in press).

'Barnstable Cl, Bodmer WF, Brown G, et al. Production ofmonoclonal antibodies to group A eythrocytes, HLA andother human ceil surface antigens-new tools for geneticanalysis. CeUl 1978;14:8-20.

20Abrahamson CS, Kersey JH, LeBien TW. A monoclonal anti-body (BA-I) reactive with cells of human B lymphocytelineage. J Immunol 1981;126:83-8.

Callard RE, Smith CM, Worman C, Linch D, Cowley JC, Bever-ley DCL. Unusual phenotype and function of an expandedsubpopulation of T cells in patients with haemopoietic disor-ders. Clin Exp Immunol 1981;43:497-505.

22 Balfour BM, Drexhage HA, Kamperdijk EWA, Hoefsmit EM.Antigen presenting cells, including Langerhans cells, veiledcells and interdigitating cells. In: Microenvironments inhaemopoietic and Iymphoid differentiation. Ciba Foundation

Symposium 84. London: Pitman Medical, 1981:281-301.3 Poulter LW, Duke 0, Hobbs S, Janossy G, Panayi G. Histochem-

ical discrimination of HLA-DR+ cell populations in the nor-mal and arthritic synovial lining. Clin Exp Immunol1982;48:381-8.

24Poulter LW, Seymour GJ, Duke 0, Janossy G. Immunohistolog-ical analysis of delayed type hypersensitivity in man. CeUlImmunol 1982;74:358-69.

25 Steinman RM, Witner MD. Lymphoid dendritic cells are potentstimulators of the primary mixed leucocyte reaction in mice.Proc Nad Acad Sci USA 1978;75:5132-6.

26 Berg, B, Ringden 0. Correlation between relative responses inmixed lymphocyte culture, HLA-D and DR typing and graftsurvival in renal transplantation. Transplantation1982;33:291-2.

Requests for reprints to: Dr MJ Raftery, Department ofNephrology and Transplantation, Royal Free HospitalLondon NW3 2QG, England.

741

on Septem

ber 4, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.36.7.734 on 1 July 1983. D

ownloaded from