Distinct regulation of HLA class II and class I cell surface expression in the THP1 macrophage cell line after bacterial phagocytosis

0014-2980/99/0202-499$17.50+.50/0© WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999

Distinct regulation of HLA class II and class I cellsurface expression in the THP-1 macrophage cellline after bacterial phagocytosis

Andrea De Lerma Barbaro1, Giovanna Tosi1, Maria Teresa Valle2, Anna MariaMegiovanni2, Silvia Sartoris3, Antonella D’Agostino1, Ornella Soro4, Maria CristinaMingari5, G. Walter Canonica6, Fabrizio Manca2 and Roberto S. Accolla1, 7

1 Unit of Cellular and Molecular Genetics, Advanced Biotechnology Center, Genova, Italy2 Servizio/Cattedra di Immunologia, Ospedale S. Martino, University of Genova, Genova, Italy3 Institute of Immunology and Infectious Diseases, School of Medicine, University of Verona,

Verona, Italy4 Institute of Microbiology, School of Medicine, University of Genova, Genova, Italy5 Department of Clinical and Experimental Oncology, School of Medicine, University of Genova,

Genova, Italy6 Department of Internal Medicine, School of Medicine, University of Genova, Genova, Italy7 Department of Clinical and Biological Sciences, Medical School, University of Insubria, Varese,

Italy

Expression of HLA and CD1b molecules was investigated in the THP-1 macrophage cell linewithin 2 weeks following phagocytosis of mycobacteria or Escherichia coli. During the first2–3 days, cell surface expression of HLA class II and CD1b was drastically down-modulated, whereas HLA class I expression was up-modulated. In the following days bothHLA class II and CD1b expression first returned to normal, then increased and finallyreturned to normal with kinetics similar to that observed for the steadily increased HLA classI. The initial down-modulation of HLA class II and CD1b cell surface antigens was absolutelydependent on phagocytosis of bacteria. Further studies indicated that initial HLA class II cellsurface down-modulation (1) was not due to reduced transcription or biosynthesis of matureHLA class II heterodimers, (2) was only partially, if at all, rescued by treatment with IFN- + ,although both mRNA and corresponding intracellular proteins increased up to sixfold withrespect to untreated cells, and (3) resulted in failure of THP-1 cells to process and presentmycobacterial antigens to HLA-DR-restricted antigen-specific T cell lines. The existence ofa transient block of transport of mature HLA class II heterodimers to the cell surface in thefirst days after phagocytosis of bacteria may have negative and positive consequences: itdecreases APC function early but it may increase it later by favoring optimal loading of bac-terial antigens in cellular compartments at high concentration of antigen-presenting mole-cules.

Key words: HLA expression / Phagocytosis / Bacterium / CD1b / IFN- +

Received 27/2/98Revised 21/10/98Accepted 22/10/98

[I 18117]

1 Introduction

Specific immunity against bacteria involves a series ofprocesses which initiate with the phagocytosis of thebacterium by appropriate phagocytic cells, mainly of themonocyte-macrophage lineage, designated APC, degra-dation of the microorganism in specific intracellular com-partments, intracellular binding of bacterial antigenic

components to specialized molecules, and exposure ofthese antigens to the cell surface where they can be rec-ognized by specific regulatory T cells, mainly of the CD4phenotype [1]. T cell recognition is required to triggerimmune effector functions such as antibody productionand cytotoxic responses. In APC, the molecules involvedin the presentation of foreign antigens to T cells aremainly represented by the very polymorphic MHCencoded class II molecules, designated in man HLA-DR,-DP and -DQ. These molecules bind antigenic peptidesin specific acidic compartments, migrate to the cell sur-face and present their bound peptides to CD4 Th cells[2]. Other polymorphic HLA-encoded structures, the

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 499

class I molecules HLA-A, B and C, are also involved inthe presentation of peptides, but in this case the pre-sented peptides are derived from endogenously synthe-sized molecules, such as viral antigens, degraded in thecytosol and captured by nascent class I molecules in theendoplasmic reticulum. After exposure to the cell sur-face, class I-bound peptides are recognized by cytotoxicT cells mainly of the CD8 phenotype [3]. Recently it hasbecome apparent that, beside HLA-class II- and class I-restricted peptide antigen presentation, other antigenpresentation pathways exist which involve non-polymorphic cell surface structures such as CD1 mole-cules [4]. CD1 binds and presents glycolipid antigens ofbacterial origin to other T cells mainly expressing the § gTCR [5, 6]. It is likely that HLA class II-restricted andCD1-mediated antigen presentations are both importantin the vast majority of infections by extracellular andintracellular bacteria. In this latter category the case ofmycobacteria is of particular interest because thesemicroorganisms infect human macrophages, which areprofessional APC. In this cellular environment mycobac-teria may persist for long time without being killed anddegraded, and actually with the capacity to divide. Thelong intracellular persistence of mycobacteria may causeleaking of bacterial antigens into the cytosol where bac-terial proteins are degraded and potentially presentedthrough the HLA class I presentation pathway [1]. Thus,both HLA class II and class I histocompatibility mole-cules can participate in presentation of mycobacterialantigens to specific T cells.

With these premises, it appeared important to us to ana-lyze the HLA class II, HLA class I and CD1 moleculeexpression as one of the key parameters for the evalua-tion of an efficient presentation of bacterial antigens tothe immune system.

We approached this aspect by studying a human macro-phage cell line, THP-1, after phagocytosis of heat-killedmycobacteria or gram-negative Escherichia coli. Heat-killed bacteria were used because in this study wewanted to focus on the effect of bacterial phagocytosisas opposed to phagocytosis of inert particles, on theexpression of antigen-presenting molecules during thefirst days after phagocytosis. Although this strategy hasthe obvious limitation of disregarding the aspect of themetabolic activity of live bacteria within the phagocyticcells, it has the advantage that the comparison betweenphagocytic processes in quantitative terms (similar num-bers of bacteria offered to the phagocyte) can be maxi-mized and possible differences in the phenotype ofphagocytes treated with gram-negative and with myco-bacteria, respectively, can be related to structural char-acteristics of the two microorganisms.

We observed interesting differences in the expression ofHLA class II with respect to class I cell surface antigens.Moreover, strong variations in the expression of CD1 cellsurface antigens were also observed which closely par-alleled the ones observed for HLA class II antigens. Nodramatic differences were observed between phagocyticcells treated with mycobacteria or with E. coli.

Some of the cellular and molecular mechanism leadingto the dramatic and concerted modulation of HLA classII and CD1 molecules with respect to HLA class I mole-cules, and the possible implications of these findings forthe homeostasis of the immune response against bacte-ria, are discussed.

2 Results

2.1 HLA class II and CD1b but not HLA class Icell surface antigens are similarly modulatedin THP-1 cells after phagocytosis ofmycobacteria or gram-negative E. coli

THP-1 cells were treated with various heat-killed bacte-ria including BCG, Mycobacterium tuberculosis andgram-negative E. coli. At different times after bacterialtreatment, the cell surface expression of HLA class I,class II and CD1b antigens was evaluated by indirectimmunofluorescence and cytofluorimetric analysis.

Fig. 1 shows the results of a representative experiment inwhich treatment with BCG and E. coli were compared.HLA class I cell surface expression steadily increasedover time in cells treated with either BCG or E. coli. Maxi-mum increase was observed at day 3 and high expres-sion was maintained up to day 8, after which the expres-sion began to decrease to reach the value of untreatedcells around day 13.

On the other hand, the kinetics of HLA class II expressionwas drastically different from the one of HLA class I, par-ticularly during the first 4 days after treatment. Indeed,both in BCG- and in E. coli-treated cells, the expressionof class II cell surface molecules dramatically decreasedduring the first 2 days. In BCG-treated cells a 55 %reduction compared to untreated cells cultured undersimilar conditions was observed. The expression of allassessable class II antigens was reduced, includingHLA-DR, DQ and DP. In E. coli-treated cells, the reduc-tion of HLA class II cell surface expression was evenmore dramatic, the HLA-DQ and HLA-DR expressionbeing decreased to 20 % and 40 %, respectively, whencompared with untreated cells. After day 2 HLA class IIexpression increased. At day 4 the HLA class II expres-sion returned to values of untreated cells and con-

500 A. De Lerma Barbaro et al. Eur. J. Immunol. 1999. 29: 499–511

Figure 1. Kinetics of HLA class I, class II and CD1b cell sur-face expression in THP-1 cells after phagocytosis of deadbacteria. The expression of the various markers listed on theright of each panel was assessed by indirect immunofluores-cence and cytofluorometry as described in Sect. 4.3. Valuesare expressed as fluorescence ratio of the value obtained incells treated with bacteria with respect to untreated cells cul-tured under the same conditions (mean fluorescence chan-nel). The fluorescence ratio was assessed at different days asindicated. Upper panel: THP-1 cells treated with BCG; lowerpanel: THP-1 cells treated with E. coli. The basic cell surfaceexpression (mean fluorescence channel) of the various mark-ers in untreated cells was as follows: HLA-A, B, C j 19.8;HLA-DR j 7.2; HLA-DQ j 6.8; CD1b j 8.9. These valuesdid not vary by more than 10 % over the kinetics period. Thenegative control (irrelevant isotype-matched antibody) wasalways less than 0.3.

tinued to increase thereafter, reaching at day 6–8expression levels 2-fold and 2.7-fold those of untreatedcells in BCG-treated and E. coli-treated cells, respec-tively. After day 8 the expression returned to normal withkinetics similar to that of HLA class I cell surface anti-gens.

Table 1. Phagocytosis of latex beads does not alter HLAclass I, HLA class II and CD1b cell surface expression inTHP-1 cells

Day

Marker 2 4 6 8 10 12

HLA class I 1.09a) 1.06 0.98 1.02 1.11 1.01

HLA-DR 0.92 1.10 1.02 1.05 0.93 1.05

HLA-DQ 1.16 0.89 0.98 1.05 1.09 1.10

CD1b 1.04 0.97 1.10 0.95 1.09 1.13

a) Values are expressed as fluorescence ratio of the valueobtained in cells treated with latex beads compared withmock-treated cells cultured under the same conditions.The expression of the various markers was assessed asdescribed in Sect. 4.3.

In similar experiments, the HLA class I and class II cellsurface expression of M. tuberculosis-treated THP-1cells closely followed the kinetics of BCG-treated cells(data not shown). Interestingly, CD1b expression wasalso strongly modulated by both BCG and E. coli and thekinetics of expression after bacterial treatment closelyfollowed the one observed for the HLA class II expres-sion. Indeed, as shown in Fig. 1, CD1b was stronglyreduced at day 2 after treatment, then graduallyincreased over time, reached normal values at day 4 andcontinued to increase to reach about threefold theexpression of untreated cells at day 8.

The modulation of HLA class I, class II and of CD1b cellsurface expression in bacteria-treated THP-1 cells corre-lated with bacterial phagocytosis and not simply withphagocytosis. As shown in Table 1, at crucial time pointsof the kinetics, the phagocytosis of inert particles suchas latex beads was ineffective in modulating either HLAclass I, HLA class II or CD1b cell surface expression.

2.2 Modulation of other THP-1 cell surfaceantigens after phagocytosis of bacteria

The THP-1 cell surface phenotypic analysis was thenextended to other markers which may be relevant forbacterial interaction and cell adhesion. Fig. 2 summa-rizes the results obtained at day 2 and 5 after BCG(upper panel) or E. coli (lower panel) treatment for CD14(the LPS receptor). CD44, and CD11a adhesion mole-cule, all expressed in substantial amounts by THP-1cells. The results related to HLA class I and class II mole-cules, and to CD1b are included for comparison. It canbe seen that the expression of CD11a was not modu-

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 501

Figure 2. Cell surface phenotype of THP-1 cells at day 2and day 5 after phagocytosis of BCG (upper panel) or E. coli(lower panel). The expression of a series of cell surface anti-gens including CD11a, CD44 and CD14 was analyzed andcompared to the expression of HLA class I, HLA class II (DR,DP and DQ), and CD1b antigens. The analysis was con-ducted by indirect immunofluorescence and flow cytofluoro-metry by using mAb specific for the various markers, fol-lowed by staining with FITC-conjugated goat F(ab)2 frag-ments specific for mouse Ig. The values are expressed aspercent increase (or decrease) of fluorescence compared tothe control value of untreated THP-1 cells.

lated at all in BCG-treated cells and only slightly up-modulated at day 5 in E. coli-treated cells. On the con-trary, the LPS receptor CD14 was strongly modulatedafter both bacterial treatments and, again, with a patternsimilar to the one observed for HLA class II cell surfacemolecules. On the other hand, the CD44 homing recep-tor molecule was not modulated at all in E. coli-treatedcells but was up-modulated at day 5 in BCG-treatedcells.

Thus, depending upon the cell surface molecule and thespecific bacterial treatment, a drastic modulation ofexpression can be generated during the first 2 weeksafter phagocytosis.

2.3 HLA class I and class II expression in THP-1cells co-cultured with bacteria-treatedhomologous cells

The different behavior of HLA class I and class II cell sur-face expression, particularly during the first 2–3 daysafter treatment with bacteria, prompted us to investigatewhether the up-modulation of class I and the down-modulation of class II were due to the action of possiblesecreted and/or diffusible products of treated cells. Co-cultures were set up in transwells in which the bottomwell was occupied by cells treated with bacteria and thetop well by control cells. Two types of controls wereused: mock-treated and latex beads-treated cells withno difference in the results obtained by using either one.The ratio between bacteria-treated and control cells wasset to 5:1 to maximize possible effects due to secretedfactors. The results shown in Fig. 3 refer to the use ofmock-treated cells as control.

With regard to HLA class I, it was found that at day 2, theup-modulation observed in treated cells (Fig. 3, “I”) wasparalleled by an up-modulation in untreated co-culturedcells (“R”) (see Fig. 3A). A similar HLA class I up-modulation for both treated and untreated co-culturedcells was observed at day 3 (Fig. 3C). As controls, similarco-cultures in which mock-treated cells (R) were seededin the bottom wells are shown in Fig. 3B and D to dem-onstrate the lack of HLA class I up-modulation inabsence of bacterial treatment.

With regard to HLA class II, it was found that the strongdown-modulation observed at day 2 (Fig. 3E) was nottransferrable to untreated co-cultured cells (R), whoseclass II expression remained unchanged as compared tonormal mock-treated cells. Moreover, at day 3 after bac-terial treatment, when class II expression began torecover but was still lower than that of untreated cells,co-cultured cells were still insensitive to down-modulation effects (see histograms I and R in Fig. 3G).

From these results we conclude that during the first fewdays after bacterial treatment up-modulation of HLAclass I cell surface antigens is due to, or at least isstrongly dependent on, secreted products of treatedcells. On the contrary, the down-modulation of HLA classII cell surface antigens does not correlate with the secre-tion of biologically active factors, but requires activephagocytosis.

502 A. De Lerma Barbaro et al. Eur. J. Immunol. 1999. 29: 499–511

Figure 3. Flow cytofluorimetric analysis of HLA class I (HLA-A, B, C, panels A–D) and HLA class II (HLA-DR, panels E–H) expres-sion in normal THP-1 cells co-cultured with THP-1 cells treated with bacteria. THP-1 cells were treated or not treated with E. colifor 16 h and then put in the bottom side of a transwell chamber. Untreated THP-1 cells were seeded in the upper well of the trans-well chamber. The proportion of treated versus untreated cells was set to 5:1. At day 2 and day 3 after bacterial or mock treat-ment both treated cells (inducer, I) and normal cells (responder, R) were removed and assayed for the expression of the relevantmarker, as described in the legend of Fig. 1. In each panel, the thick line depicts the fluorescence of the upper well respondernormal cells; the thin line depicts the fluorescence of the bottom well bacteria-treated (A, C, E, G) or mock-treated (B, D, F, H)cells. Fluorescence values are expressed as log fluorescence in arbitrary units (a.u.).

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 503

Figure 4. Expression of HLA class II cell surface antigens inTHP-1 cells after phagocytosis of bacteria and treatmentwith IFN- + . Cells were treated for 14 h with either BCG, E.coli or latex beads (Beads) in presence ( | ) or in absence ( ß )of IFN- + , washed extensively and further cultured for addi-tional 34 h with the cytokine ( | ) or with medium ( ß ). HLA-DR and HLA-DQ cell surface phenotype was assessed byindirect immunofluorescence and flow cytometry asdescribed in Sect. 4.3. Columns represent the fluorescencevalues, expressed as mean fluorescence channel.

2.4 Effect of IFN- q on HLA class II expression ofTHP-1 cells treated with bacteria: phenotypicand molecular analysis

IFN- + is one of the most potent inducers of HLA class IIexpression in macrophages and other APC. It was there-fore important to investigate whether treatment of THP-1cells with this cytokine could overcome the negativeeffect on HLA class II cell surface expression observedduring the first 2 days after phagocytosis of bacteria. Tothis end, cells were incubated with or without bacteria inthe presence of the cytokine, washed and put in culturewith complete medium supplemented with recombinantIFN- + . At day 2 cells were harvested and analyzed fortheir HLA class II cell surface phenotype. Fig. 4 showsthe results of a representative experiment in which HLA-DR and HLA-DQ cell surface expression were assessed.It can be seen that the expression of both HLA-DR and-DQ molecules was strongly increased in latex beads-treated cells after incubation with the cytokine. In BCG-treated cells IFN- + increased the DR expression but onlyto 57 % of the expression values obtained in cells treatedwith latex beads. In E. coli-treated cells, incubation withthe lymphokine was dramatically less efficient in rescu-ing DR expression after bacterial treatment. Indeed IFN-

+ could only increase the expression of DR to values sim-ilar to those of latex beads-treated cells not incubatedwith the cytokine. The apparent incapacity of IFN- + toovercome the down-regulation of HLA class II in cellstreated with bacteria was even more evident for the HLA-DQ subset. In this case the cytokine could rescue DQexpression in BCG-treated cells only up to 46 % of thevalues obtained in the corresponding controls treatedwith latex beads and IFN- + , and was totally ineffective inrescuing the almost abrogated DQ expression in E. coli-treated cells.

The down-modulation of HLA class II cell surfaceexpression in the first 2 days after bacterial treatmentand the impossibility to fully rescue this expression byIFN- + prompted us to investigate in more detail themolecular correlate of this event. HLA class II mRNAanalysis and quantitation of the total amount of matureclass II heterodimers were performed. Fig. 5 shows theNorthern blot of HLA class II DR g mRNA (upper panel)and the Western blot of total HLA-DR molecules (lowerpanel). To better appreciate differences between sam-ples, two dilutions of mRNA and of protein extracts arepresented for each sample. Equivalent amounts ofmRNA and protein extract from the class II-positiveB cell line Raji are included for comparison. The resultspresented refer to THP-1 cells treated with E. coli. Quali-tatively similar results were obtained with cells treatedwith BCG (data not shown).

It can be seen that at the time of maximal down-modulation of HLA class II cell surface antigens followingbacterial treatment [THP-1 (bact)], the total amount ofDR g mRNA was similar with respect to mock-treated(THP-1) cells. THP-1 cells treated with bacteria and incu-bated with IFN- + [THP-1 (bact) + IFN- + ] stronglyincreased the transcription of DR g mRNA compared toboth their counterpart not incubated with the lymphokine(sixfold) and the mock-treated cells. However, thisincrease was twofold lower than the correspondingincrease observed in THP-1 cells not treated with bacte-ria and incubated with the lymphokine [compare THP-1(bact) + IFN- + with THP-1 + IFN- + ]. Results virtuallysuperimposable to those of the DR g mRNA wereobtained with DR § , as well as with DQ § and DQ g spe-cific mRNA (data not shown).

These results indicate that the dramatic but transientdown-modulation of HLA class II cell surface antigensafter bacterial phagocytosis, even in presence of IFN- + ,is not due to a transcriptional defect.

At the protein level, the total amount of assessible classII molecules in the various samples correlated with thecorresponding amount of class II mRNA. Interestingly, as

504 A. De Lerma Barbaro et al. Eur. J. Immunol. 1999. 29: 499–511

Figure 5. Assessment of HLA-DR-specific transcripts andcorresponding mature molecules in THP-1 cells after phago-cytosis of bacteria and treatment with IFN- + . The analysiswas conducted on the cells whose phenotype is describedin Fig. 4, and for bacteria-treated cells, only the resultsrelated to the use of E. coli are reported. (A) Northern blotanalysis of DR- g mRNA and, as a control, of g -actin. (B)Western blot of total HLA-DR molecules; similar loading ofprotein was verified after electroblot by staining the filterwith Ponceau Red, Raji, HLA class II-positive B cell line;RJ 2.2.5, HLA class II-negative derivative of Raji cells. Foreach THP-1 sample, two dilutions of total cytoplasmic RNAor protein (1:1 and 1:4) were used to maximize comparisonbetween different samples. Abbreviations used: THP-1, nor-mal cells; THP-1 + IFN- + , cells treated with IFN- + ; THP-1(bact), cells treated with bacteria; THP-1 (bact) + IFN- + , cellstreated with bacteria and incubated with IFN- + .

shown in Fig. 5B, in cells treated with bacteria the totalamount of HLA-DR molecules was not only comparablebut even higher than that of mock-treated cells [seeTHP-1 (bact) and THP-1 lanes, respectively]. Bacteria-treated THP-1 cells incubated with IFN- + [THP-1 (bact) +IFN- + ] strongly increased their content of class II mole-cules. The amount of HLA-DR proteins in bacteria- andIFN- + -treated cells, although not as high as thatexpressed in THP-1 cells treated only with IFN- + , wasfour- to fivefold higher than the corresponding amountexpressed in normal THP-1 cells, and yet the levels ofclass II cell surface antigens were similar to, if not lowerthan, those found in normal THP-1 cells (see Fig. 4). It

must be stressed that the mAb used to detect HLA-DRantigens (D1–12) reacts with a conformational determi-nant present on assembled § g heterodimers and prefer-entially with heterodimers complexed with antigenicpeptides [7, 8]. Thus the DR molecules shown in Fig. 5represent mostly terminally mature products.

Taken together, these data indicate that IFN- + rescuesboth transcription and translation in bacteria-treatedcells. However, this is not sufficient to comparativelyincrease the level of cell surface expression of the classII antigens. Thus a strong reduction, if not a block, intransport of newly synthesized mature heterodimers tothe cell surface exists during the first 2 days after bacte-rial phagocytosis.

2.5 Antigen processing and presentationcapacity of THP-1 cells during the HLA classII down-modulation induced by bacterialphagocytosis

The strong down-modulation of HLA class II cell surfaceantigens during the first 2 days after bacterial phagocy-tosis, both in unstimulated and IFN- + -induced THP-1cells, prompted us to investigate whether this could havean effect on processing and presentation of soluble pro-tein antigens to HLA-DR-restricted T cells. Fig. 6 showsthe results of a typical experiment in which THP-1 cellsafter down-modulation of HLA class II, at 48 h after bac-terial treatment (the same cells whose class II phenotypeis shown in Fig. 4), were used as APC. Ag85, a highlypurified mycobacterial secretory protein, and its immu-nodominant peptide pep11 were used as antigens andCD4+, HLA-DR-restricted, pep11-specific T cells asresponder cells. It can be seen that both mock- and latexbead-treated THP-1 cells process Ag85 and present therelevant peptide very poorly. Thus, possible functionaldifferences associated with down-modulation of theconstitutive class II expression induced by bacterialphagocytosis (in this case E. coli) in THP-1 cells couldnot be appreciated by the T cell proliferation assay used.On the other hand, incubation with IFN- + resulted in astrong induction of both Ag85 processing and pep11presentation to pep11-specific T cells in normal and inlatex bead-treated THP-1 cells. This was paralleled bythe strong increase in HLA class II cell surface expres-sion induced by the cytokine (see Fig. 4). However, incu-bation with IFN- + of E. coli-treated THP-1 cells did notresult in appreciable antigen processing and presenta-tion capacity and this was paralleled by the profounddown-modulation of HLA class II cell surface antigensobserved in E. coli-treated cell even after incubation withthe cytokine.

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 505

Figure 6. HLA-DR-restricted, peptide-specific T cell prolif-eration in response to antigen presentation by THP-1 cells. ACD4+ T cell line specific for peptide 11 (pep11) of the myco-bacterial secretory protein Ag85 was used. THP-1 cells afterdifferent treatments were used as APC: mock, untreated;IFN- + , THP-1 cells incubated for 48 h with 500 U/ml IFN- + ;E. coli, THP-1 cells at 48 h after phagocytosis of E. coli;E. coli + IFN- + , THP-1 at 48 h after phagocytosis and cul-tured in continuous presence of IFN- + ; beads, THP-1 cells at48 h after phagocytosis of latex beads; beads + IFN- + , THP-1 at 48 h after phagocytosis of latex beads and cultured incontinuous presence of IFN- + . Values are expressed asstimulation index (SI), i.e. the ratio between the CD4+ T cellproliferation levels in presence and in absence of APC.White and black columns represent the SI obtained inresponse to APC incubated with the peptide pep11 or withthe protein Ag85, respectively.

3 Discussion

Previous studies concerning the expression of cell sur-face structures involved in antigen presentation afterphagocytosis of bacteria and/or infection of monocytes,particularly with mycobacteria, are scanty and often con-tradictory. A down-modulation of HLA class II expressionin the 16 h following phagocytosis of both mycobacteriaand other bacteria were reported by Flad and associates[9, 10], whereas others found no modulation of HLA-DRafter phagocytosis of M. avium-M. intracellulare [11].

Moreover, it is not clear whether the expression of cellsurface antigens, including HLA class I and class II mole-cules, can be differentially modulated depending on thephagocytosis of distinct bacteria or even of the samemicroorganism in either viable or heat-killed form [9–14].

Beside the different methodological approaches, anadditional difficulty in interpreting the available datastems from the poor information on the kinetics of varia-tion in HLA cell surface markers during time after initialphagocytosis, and on comparative kinetics of expressionof HLA- and non-HLA-encoded cell surface structuresinvolved in presentation of bacterial antigens to T cells.

In this study we have undertaken a systematic analysisof HLA expression in a cellular system, the THP-1 mono-cyte cell line, whose phenotype closely resembles that ofa normal macrophage. THP-1 cells are very active inphagocytic function. In this system we have analyzed thephenotypic changes occurring after phagocytosis ofeither gram-negative E. coli or M. bovis (BCG) and M.tuberculosis during a period of 13 days.

Interestingly, a completely different behavior of cell sur-face modulation of HLA class I as compared to class IIantigens was observed. Indeed, HLA class I expressionsteadily increased from the very beginning after phago-cytosis to day 6–8, and then returned to levels ofuntreated cells at the end of the 13-day period. Con-versely, HLA class II cell surface expression dramaticallydecreased during the first 2 days after phagocytosis,then returned to normal levels around day 4, and contin-ued to increase up to day 6–7, to finally follow a kineticssimilar to that of class I cell surface antigens. The differ-ences observed in THP-1 cells treated with mycobacte-ria as compared to cells treated with E. coli were morequantitative than qualitative, suggesting that, in the caseof phagocytosis of dead bacteria, the modulation ofmacrophage HLA cell surface antigens was a commonevent irrespective of the microorganism internalized.

It was important to establish the molecular correlate ofthe transient HLA class II cell surface down-modulationobserved during the first 2–3 days after bacterial phago-cytosis. The presence of soluble factors of either bacte-rial or cellular origin with suppressing characteristics wasexcluded by co-culture experiments of bacteria-treatedand untreated cells. Indeed, in the latter cells the expres-sion of HLA class II cell surface molecules was unaf-fected, whereas the expression of HLA class I molecules(see below) could be up-modulated. We then askedwhether reduced biosynthesis of class II moleculescould be responsible for the down-modulation observed.Analysis of specific mRNA and of the total pool of classII molecules demonstrated that there was no appreciablereduction in the amount of mature transcripts and, moreimportantly, no reduction in the amount of mature HLAclass II heterodimers. Treatment with IFN- + of the cellsundergoing phagocytosis of bacteria further reinforced inthe notion that HLA class II down-modulation was notdue to reduced biosynthesis. In fact, IFN- + was able to

506 A. De Lerma Barbaro et al. Eur. J. Immunol. 1999. 29: 499–511

drastically increase the transcription of class II genesand the synthesis of mature class II heterodimers by six-to sevenfold in bacteria-treated cells, and yet no compa-rable drastic increase in the amount of cell surface classII molecules was observed. Indeed, in E. coli- plus IFN-+ -treated cells the amount of class II cell surface anti-gens was even lower than that observed in normal cellstreated neither with bacteria nor with the cytokine. Theexperiments with IFN- + indicated, however, that class IItranscription could be partially inhibited in induced cellsafter phagocytosis of bacteria. Indeed, bacteria-treatedcells incubated with the cytokine showed at least a two-fold reduction of both class II mRNA and correspondingprotein with respect to mock-treated cells incubatedwith IFN- + .

Taken together, these results strongly suggest thatmechanisms of blockade or inhibition of the transport ofnewly synthesized mature class II heterodimers to thecell surface can be responsible for the observed tran-sient down-modulation of HLA class II cell surface anti-gens. Internalization and segregation of cell surface mol-ecules in specific endocytic compartments can also par-ticipate in the observed HLA class II down-modulationafter bacterial phagocytosis. Experimental evidence ofHLA class II internalization and recycling to the cell sur-face has been obtained [15]. Under certain circum-stances, and relevant to the present study, internalizationis not followed by recycling, as observed during phago-cytosis of live and, to a lesser extent, dead M. tuberculo-sis [14], as well as Leishmania amazonensis amastigotes[16].

A recent investigation [17] has shown that incubation ofhuman monocytes with IL-10 results in class II cell sur-face down-regulation, which is apparently due to both ablock in exocytosis of mature molecules to the cell sur-face and to a reduced recycling of cell surface mole-cules. This effect can be seen on both the constitutiveand IFN- + -induced class II expression and is clearlyappreciated within 24 h after IL-10 treatment, althoughthe kinetics were not investigated further [17]. Theseresults are strikingly similar to the results presented here;however, they are clearly distinct by the causative event,since in our case phagocytosis of bacteria is requiredand soluble factors cannot substitute for it. Moreover,bacteria are not continuously present during the assayas IL-10 is in the above studies. Thus, if internalization ofHLA class II molecules not followed by recycling takesplace, it should be limited to the earlier phases of thekinetics when phagocytosis is still ongoing. Neverthelessit is of importance that diverse stimuli which can be seenas immunomodulatory in nature can transiently but dra-matically affect the post-synthetic steps involved in thecell surface expression of HLA class II molecules. Stud-

ies are now in progress to determine the contribution ofthe various post-synthetic steps (transport, internaliza-tion, recycling) to the transient HLA class II down-modulation after bacterial phagocytosis.

Whatever the mechanism, class II cell surface down-modulation has important functional consequences onthe antigen processing and presentation capacities ofTHP-1 cells. Although THP-1 cells do not normally exertappreciable APC function, they can be induced to do soby treatment with IFN- + . However, treatment with thiscytokine did not rescue antigen processing and/or pres-entation function in THP-1 cells at the time of maximumHLA class II down-modulation after phagocytosis. Theseresults reinforce the notion that a correct quantitativeexpression of cell surface MHC class II molecules isrequired to accomplish APC functions. Moreover, if theseresults reflect a similar behavior of macrophages in vivoduring bacterial infections, it is possible that during thefirst days after phagocytosis macrophages are function-ally unable to exert correct APC functions, even in pres-ence of strong inflammatory stimuli like IFN- + .

The particular behavior observed for HLA class II cell sur-face antigens was shared by other non-HLA-encodedcell surface receptors which are involved in binding ofbacterial products and internalization, and in presenta-tion of bacterial antigens to T cells. This was the case forCD1b molecules, which are constitutively expressed inTHP-1 cells, and strongly implicated in the presentationof mycobacterial components such as lipoarabinoman-nan (LAM) and mycolic acid to T cells [5, 6]. It was alsothe case for CD14, the LPS receptor, which also bindsmycobacterium-derived LAM [18].

It has been recently demonstrated that CD1b moleculesco-localize with MHC class II molecules in the sameintracellular compartments where class II molecules arelikely loaded with antigenic peptides [19]. It has alsobeen shown that a significant proportion of these CD1bmolecules are derived from the internalization of cell sur-face molecules.

Thus, on the basis of the results presented in this studyand on previous observations, it is tempting to speculatethat the cell surface expression of MHC class II andCD1b molecules is tightly co-regulated after bacterialphagocytosis and finalized to focus both molecules inthe same intracellular compartments where they aremore prone to capture bacterial antigens. These com-partments are the ones where newly synthesized class IIand, perhaps, CD1b molecules are also focussed duringbiosynthesis, and, in the case of bacterial phagocytosis,they may accumulate and be forced to stay there for atleast 48 h. The fact that the cell surface expression of

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 507

both HLA class II and CD1b molecules increases after 4days from phagocytosis implies that the block in recy-cling and/or in the transport to the cell surface can bereleased, and studies are in progress to investigate thisparticular aspect. All these events may contribute to anoptimal antigenic loading of HLA class II and CD1b mol-ecules that can then be transported to the cell surface ingreater amounts to better accomplish specific functionssuch as optimal and adequate bacterial antigen presen-tation to T cells.

Within this context, the similar pattern of modulation ofthe LPS receptor CD14, which can also bind LAM andthus favor both E. coli and BCG internalization [18], maybear importance in a possible integrated function of sev-eral cell surface receptors involved in bacterial recogni-tion and specific antigen presentation.

In contrast to HLA class II and CD1b, the level of HLAclass I cell surface molecules steadily increased afterbacterial phagocytosis. This event correlated with thepresence of soluble factors that could up-regulate theHLA class I expression of normal THP-1 cells in a similarfashion. Whether these factors are bacterial productsderived from the degradation of internalized microorgan-isms or newly synthesized cellular factors in response tobacterial phagocytosis is at present under investigation.However, the persistent up-regulation of HLA class I anti-gens over a period of 10 days favors the idea that thesoluble factors are of cellular origin. This result may haveimportant functional implications. It is widely acceptedthat bacterial infections can evoke cytolytic T cellresponses mediated by CD8-positive cells [20]. This isparticularly important in those cases in which intracellu-lar pathogens are involved. Indeed, it has been shownthat g 2-microglobulin gene knockout mice which cannotexpress MHC class I antigens and, by consequence,have a deficient CD8 T cell compartment, display a verysevere form of tuberculosis when infected by a humanstrain of M. tuberculosis [21].

The continuous and persistent increase in HLA class Icell surface expression may be advantageous for animmune response if bacterial antigens, and particularlymycobacteria-derived peptides, can be released or leakoff the phagosome in which the mycobacterium istrapped. These proteins, reaching the cytosol, could bedegraded and the corresponding peptides transportedinto the endoplasmic reticulum by a mechanism depend-ing on the transporter associated with antigen presenta-tion (TAP) [3] to be captured by class I molecules andpresented to class I-restricted T cells. This, in turn, wouldfavor a cytotoxic response against macrophagesinfected with bacteria.

4 Materials and methods

4.1 Bacterial preparations

In most experiments M. bovis (BCG, from Pasteur Merieux,France) and an isolate of M. tuberculosis derived from apatient with active pulmonary tuberculosis were used. Asgram-negative bacteria, the DH5- § strain of E. coli wasused. Mycobacteria were grown in Löwenstein-Jensen solidmedium for 15–20 days, harvested, carefully resuspended inPBS and gently sonicated and centrifuged at low speed toremove clumps. E. coli were grown in LB medium overnight,washed and resuspended in PBS. Aliquots of both bacteriawere stored at −80 °C before use.

Mycobacteria were killed by autoclave treatment for 15 minwhereas E. coli were treated for 30 min at 72 °C.

4.2 Cells

The human monocyte-macrophage THP-1 cell line wasused in the present study. THP-1 cells express macrophagemarkers such as CD14; they express both HLA class I andclass II cell surface antigens. In this respect they resembleactivated pulmonary macrophages which express HLA classII, in contrast to resting macrophages that do not expressthese HLA markers. Cells were grown under standard cul-ture conditions in RPMI 1610 medium with 10 % FCS with-out antibiotics.

4.3 Cell surface phenotype and monoclonal antibodies

Cell surface phenotype was assessed by indirect immuno-fluorescence and cytofluorometry by using mAb specific forthe various HLA class I and class II antigens, for CD1b mole-cules, and for a series of other cell surface antigens includ-ing CD14, CD44, and CD11a.

The following reagents were used: B9.12.1, specific formonomorphic determinants of all HLA class I molecules;D1.12 and B7-21, specific for HLA-DR and -DP class IImonomorphic determinants, respectively; XIII 358.4, spe-cific for HLA-DQ class II molecules expressed in the DQ2haplotypes; CD1b (mAb 4A7-6.5), CD14 (mAb 10.43.3H7),CD11a (mAb 8F2.7), CD44 (mAb 8.B2.5) specific antibodieswere kindly donated by Dr. Daniel Olive (Marseille). Cytofluo-rometric analysis was performed as described [22]. Briefly,cells were collected during the exponential phase of growth,washed with cold RPMI medium and resuspended at a con-centration of 1 × 106 cells/ml. Hundred microliters of cellsuspension were incubated with saturating concentrationsof specific mAb for 1 h at 4 °C. After washing, cells wereincubated with purified FITC-labeled goat F(ab)2 fragmentsspecific for mouse Ig for 1 h and then analyzed on a EpicsProfile cytofluorometer (Coulter Corp., Hialeah, USA).

508 A. De Lerma Barbaro et al. Eur. J. Immunol. 1999. 29: 499–511

4.4 Evaluation of the effects of bacterial treatment onTHP-1 cells

THP-1 cells in exponential phase of growth were centri-fuged, washed in RPMI medium and resuspended at a con-centration of 2 × 106/ml in complete culture medium. Heat-killed bacteria were then added to the cells, usually at50–100:1 bacteria :cell ratio. The mixture was incubated for2 h at 37 °C under agitation and then for 14 h in the incuba-tor at the same temperature. The mixture was then centri-fuged at low speed to separate cells from bacteria. Cell pel-lets were washed three times with warmed medium andresuspended in RPMI-10 % FCS at 2 × 105 cells/ml. Prelimi-nary experiments showed that at this time the vast majority(between 80 and 95 %) of THP-1 cells had phagocytosedbacteria; the number of particles per cell varied between 2 to15 as assessed by Ziehl-Neelsen stain. At different timesafter bacterial treatment (see Sect. 2) cells were collectedand their cell surface phenotype analyzed as describedabove. In parallel experiments, the effect on cell surfacephenotype of cells undergoing phagocytosis of inert latexbeads (2 ? m diameter; Coulter Electronics, Luton, GB) wasevaluated.

To assess the production of cellular factors with possibleactivatory or inhibitory functions on cell surface markerexpression, co-culture experiments were carried out in atrans-well culture plate (Costar, Milan, Italy) allowing freeexchange of culture medium and solutes. Treated anduntreated cells were seeded in the lower and upper well,respectively, at a 5:1 ratio. At given times both treated (stim-ulator) and untreated (responder) cells were collected andthe cell surface phenotype assessed.

4.5 HLA class II cell surface and molecular analysis ofTHP-1 cells treated with bacteria and IFN- q

The effect of bacterial phagocytosis on the IFN- + -mediatedup-regulation of HLA class II expression was also evaluated.For this experiment, THP-1 cells undergoing phagocytosisof bacteria were incubated with 500 U/ml IFN- + at the begin-ning of bacterial treatment. After incubation with bacteriaand washing (see above), cells were put in culture and sup-plemented again with the same concentration of the cyto-kine, and analyzed after 2 days of culture for their cell sur-face phenotype by indirect immunofluorescence and cyto-fluorimetry.

At the same time cells were also analyzed for their total con-tent of mature class II heterodimers by Western blot, and fortheir class II-specific mRNA by Northern blot.

For protein analysis, 1 × 106 cells were lysed in 100 ? l oflysis buffer containing 50 mM Tris, 150 mM NaCl, 1 mMEDTA, 1 % NP40, 1 mM phenylmethylsulfonyl fluoride,20 ? M aprotinin, 20 ? M leupeptin, pH 7.5. Thirty microlitersof cell lysate were electrophoresed on SDS-12 % polyacryl-

amide minigels and then blotted on polyvinylidene difluoridetransfer membranes (Biotechnology Systems, NEN Du Pont,Milan, Italy). Filters were incubated with anti-HLA-DR or -DQantibodes, washed and further incubated with rabbit anti-bodies (IgG fraction) specific for mouse IgG (Sigma Immu-nochemicals, Milan, Italy). Filters were developed with thechemiluminescence method using Renaissance WesternBlot reagent (New England Nuclear, Life Science Products,Milan, Italy) and exposed on Amersham Hyperfilm films.

Cytoplasmic RNA was prepared following standard proce-dures [23], electrophoresed in 1.5 % agarose gel containing2.2 M formaldehyde and transferred to N-Hybond nylonmembranes in 20 × SSC. HLA class II-specific probes wereexcised from plasmids and purified as previously described[24]. The various purified probes were uniformly labeled byrandom priming (Random-Primed DNA labeling kit, Boehrin-ger Mannheim Italia, Milan, Italy) and added to hybridizingfilters at a concentration of 3 × 106 cpm/ml. Prehybridizationand hybridization were performed in Amersham RapidHybridization Buffer, following the manufacturer’s recom-mendations. Filters were autoradiographed for 24–48 h,using Amersham Hyperfilm films.

4.6 Antigen processing and presentation to T cells

THP-1 cells that had phagocytosed bacteria, latex beads ornothing were analyzed for their antigen processing and pres-entation capacity at the time of maximum down-modulationof HLA class II antigens. Briefly, after 48 h from bacterialtreatment cells were collected, washed, irradiated at 6000rad with a 137Cs source (Gammacell, Nordion, Inc., Kanata,Canada) and dispensed in a flat-bottom microtiter plate at aconcentration of 1 × 104 cells/well, based on preliminarytitration experiments for optimal APC/T cell responder ratio(see below). Cell were then pulsed with 5 ? g/ml Ag85, ahighly purified mycobacterial secretory protein [25] or with1 ? g/ml of a peptide of AG85, designated pep11 (aminoacids 91–108, GCQTYKWETLLTSELPQW), for 4 h. CD4+

T cells (2 × 104 from an established line specific for pep11and restricted by the HLA-DR2 allele expressed in THP-1cells) were then added to the wells, as described [26, 27].After 2 days the cultures were pulsed with 0.5 ? Ci [3H]thymi-dine (5 Ci/mmol specific activity, Amersham, Amersham,GB) and harvested 12 h later with a Filtermate 196 appara-tus. The dried filters were counted in a Matrix 96 counter(Packard Instruments, Downers Grove, IL).

Acknowledgements: R. S. Accolla wishes to thank Prof.Lorenzo Moretta, IST, University of Genova, for advise andcontinuous support of his research. This work was sup-ported by the following grants to R.S.A.: Istituto Superiore diSanita (I.S.S.) National Project “Tuberculosis”; I.S.S.National Research Project on A.I.D.S. no. 40A.0.01; MURSTNational Project “Meccanismi di resistenza alle neoplasie”;Consiglio Nazionale delle Ricerche (CNR) Target Project on

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 509

“Biotechnology”; and the following grants to F.M.: I.S.S.,National Project “Tuberculosis”, I.S.S. National ResearchProject on AIDS, no. 40A.0.64, and CNR Target Project on“Biotechnology“.

References

1 Kaufmann, S. H. E., Immunity to intracellular bacteria.Annu. Rev. Immunol. 1993. 11: 129–163.

2 Germain, R. N. and Margulies, D. H., The biochemistryand cellular biology of antigen processing and presenta-tion. Annu. Rev. Immunol. 1993. 11: 403–450.

3 Cerundolo, V. and Braud, V., Cell biology of MHC classI molecules. In Browning, M. and McMichael, A. (Eds.),HLA and MHC: Genes, Molecules and Function. BIOSScientific Publishers, Oxford 1996, pp 193–223.

4 Porcelli, S. A., The CD1 family: a third lineage ofantigen-presenting molecules. Adv. Immunol. 1995. 59:1–98.

5 Beckman, E. M., Porcelli, S. A., Morita, C. T., Behar, S.M., Furlong, S. T. and Brenner, M. B., Recognition of alipid antigen by CD1-restricted § g + T cells. Nature 1994.372: 691–694.

6 Sieling, P. A., Chatterjee, D., Porcelli, S. A., Prigozy, T.I., Soriano, T., Brenner, M. B., Kronenberg, M., Bren-nan, P. J. and Modlin, R. L., CD1-restricted T cell rec-ognition of microbial lipoglycans. Science 1995. 269:227–230.

7 Accolla, R. S., Gross, N., Carrel, S. and Corte, G., Dis-tinct forms of both § and g subunits are present in thehuman Ia molecular pool. Proc. Natl. Acad. Sci. USA1981. 78: 4549–4551.

8 Roucard, C., Garban, F., Mooney, N. A., Charron, D. J.and Ericson, M. L., Conformation of human leukocyteantigens class II molecules. Evidence for superdimersand empty molecules on human antigen presentingcells. J. Biol. Chem. 1996. 271: 13993–14000.

9 Pryma, J., Baran, J., Ernst, M., Woloszyn, M. andFlad, H. D., Altered antigen presenting capacity ofhuman monocytes after phagocytosis of bacteria. Infect.Immun. 1994. 62: 1961–1967.

10 Gercken, J., Pryjma, J., Ernest, M. and Flad, H. D.,Defective antigen presentation by Mycobacteriumtuberculosis-infected monocytes. Infect. Immun. 1994.62: 3472–3478.

11 Tsuyuguchi, I., Kawasumi, H., Takashima, T., Tsuyu-guchi, T. and Kishimoto, S., Mycobacterium avium-Mycobacterium intracellulare complex-induced sup-pression of T-cell proliferation in vitro by regulation ofmonocyte accessory cell activity. Infect. Immun. 1990.58: 1369–1378.

12 Kaye, P. M., Sims, M. and Feldmann, M., Regulation ofmacrophage accessory cell activity by mycobacteria. II.In vitro inhibition of Ia expression by Mycobacterium mi-croti. Clin. Exp. Immunol. 1986. 64: 28–34.

13 Clemens, D. L. and Horwitz, M. A., Membrane sortingduring phagocytosis: selective exclusion of Major Histo-compatibility Complex molecules but not complementreceptor CR3 during conventional and coiling phagocy-tosis. J. Exp. Med. 1992. 175: 1317–1326.

14 Clemens, D. L. and Horwitz, M. A., Characterization ofthe Mycobacterium tuberculosis phagosome and evi-dence that phagosomal maturation is inhibited. J. Exp.Med. 1995. 181: 257–270.

15 Pinet, V., Vergelli, M., Martin, R., Bakke, O. and Long,E. O., Antigen presentation mediated by recycling of sur-face HLA-DR molecules. Nature 1995. 375: 603–606.

16 De Souza Leao, S., Lang, T. H., Prina, E., Hellio, R.and Antoine, J.-C., Intracellular Leishmania amastigo-tes internalize and degrade MHC class II molecules oftheir host cells. J. Cell Sci. 1995. 108: 3219–3231.

17 Koppelman, B., Neefjes, J. J., deVries, J. E. andde Waal Malefyt, R., Interleukin-10 down-regulatesMHC class II § g peptide complexes at the plasma mem-brane of monocytes by affecting arrival and recycling.Immunity 1997. 7: 861–871.

18 Porcelli, S. A., Morita, C. T. and Modlin, R. L., T cellrecognition of non-peptide antigens. Curr. Opin. Immu-nol. 1996. 8: 510–516.

19 Prigozy, T. I., Sieling, P. A., Clemens, D. L., Stewart, P.,Behar, S. M., Porcelli, S. A., Brenner, M. B., Modlin, R.L. and Kronenberg, M., The mannose receptor deliverslipoglycan antigens to endosomes for presentation toT cells by CD1b molecules. Immunity 1997. 6: 187–197.

20 Reimann, J. and Kaufmann, S. H. E., Alternative anti-gen processing pathways in anti-infective immunity.Curr. Opin. Immunol. 1997. 9: 462–469.

21 Flinn, J. L., Goldstein, M. M., Triebold, K. J., Koller, B.and Bloom, B. R., Major histocompatibility complexclass I-restricted T cells are required for resistance toMycobacterium tuberculosis infection. Proc. Natl. Acad.Sci. USA 1992. 89: 12013–12017.

22 De Lerma Barbaro, A., Sartoris, B., Tosi, G., Nicolis,M. and Accolla, R. S., Evidence for a specific post-transcriptional mechanism controlling expression ofHLA-DQ, but not DR and DP, molecules. J. Immunol.1994. 153: 4530–4538.

23 Maniatis, T., Fritsch, E. F. and Sambrook, J. (Eds.),Molecular Cloning: A Laboratory Manual. Cold SpringHarbor Laboratory Press, Cold Spring Harbor 1982.

510 A. De Lerma Barbaro et al. Eur. J. Immunol. 1999. 29: 499–511

24 Sartoris, S., Scupoli, M. T., Scarpellino, L. Paiola, F.,Jotterand-Bellomo, M., Tridente, G. and Accolla, R.S., Inducible and constitutive MHC class II gene expres-sion. Distinct tissue-specific genetic controls. J. Immu-nol. 1990. 145: 1960–1967.

25 De Bruin, J., Huygen, K., Bosmans, R., Fauville, M.,Lippens, R., Van Vooren, J.-P., Falmagne, P., Weck,M., Wiker, H. G., Harboe, M. and Turneer, M., Purifica-tion, characterization and identification of a 32 kDa pro-tein antigen of M. bovis BCG. Microb. Pathog. 1987. 2:351–366.

26 Manca, F., Rossi, G., Valle, M., Lantero, S., Damiani,G., Li Pira, G. and Celada, F., Limited clonal heteroge-neity of antigen specific T cells focussing in the pleuralspace during mycobacterial infection. Infect. Immun.1991. 59: 503–513.

27 Manca, F., Fenoglio, D., Li Pira, G., Kunkl, A. andCelada, F., Effect of antigen antibody ratio on macro-phage uptake, processing and presentation to T cells ofantigen complexed with polyclonal antibodies. J. Exp.Med. 1991. 173: 37–48.

Correspondence: Roberto S. Accolla, Unit of Cellular andMolecular Genetics, Advanced Biotechnology Center, LargoRosanna Benzi, 10, I-16132 Genova, ItalyFax: +39-10-5 73 73 80e-mail: [email protected]

Eur. J. Immunol. 1999. 29: 499–511 Modulation of HLA surface expression after bacterial phagocytosis 511