Altered expression of CD43-hexasaccharide isoform on peripheral T lymphocytes from HIV-infected individuals

Objective: To examine if peripheral T lymphocytes from HIV-infected individuals show abnormalities in the surface expression of CD43, the major sialoglycoprotein of leukocytes.

Design: A series of 86 HIV-positive individuals was studied. The subjects, grouped by their peripheral CD4 cell count, were in different stages of the disease as defined by the Centers for Disease Control and Prevention (CDC).

Cited Here...: Peripheral leukocytes and isolated lymphocytes were examined by double and triple immunofluorescence flow cytometric and Western blot analyses with monoclonal antibodies, which discriminate between CD43 isoforms.

Cited Here...: We found elevated percentages of the surface expression of CD43-hexasaccharide isoform on T lymphocytes from 82 out of 86 individuals tested. Increasing percentages are progressively found in CDC groups 1, 2 and 3 patients. The expression of the molecule is remarkably biased towards the CD8 cell subpopulation. The percentage of cells bearing human leukocyte antigen-DR locus molecules (HLA-DR) is also augmented. Two subsets expressing T305 have been identified: a minor subset that co-expresses HLA-DR and T305; and a second population formed by the majority of T305-positive cells, which lack surface HLA-DR. Finally, we found CD43 bands with altered electrophoretic mobility in lysates from peripheral lymphocytes from all HIV-positive individuals tested.

Conclusion: The augmented expression of CD43-hexasaccharides and the observed cellular distribution suggest an important regulatory role for this molecule in HIV-specific responses.

Introduction

In AIDS, abnormalities related to the expression of CD43 have been described. CD43 is the major sialoglycoprotein expressed on all cells of haematopoietic origin ^[1,2]. The molecule is heavily glycosylated and the two known CD43 isoforms are generated by alternative glycosylation ^[2,3]. The first isoform binds tetrasaccharide O-glycans that generate a molecule of 115 000 M_r expressed on resting lymphocytes and monocytes. Hexasaccharides are bound to the second isoform of 130 000 M_r, expressed on neutrophils and platelets. Increasing evidence suggests that each CD43 isoform plays a distinct functional role.

CD43 is an activation receptor in immune cells ^[4-6], but its main functions are related to the regulation of cell-cell interactions. The abnormal expression of CD43 is observed in certain human pathologies, including leukemic T cells ^[7], non-haematopoietic tumours ^[8], as well as in the primary immunodeficiency Wiskott-Aldrich syndrome ^[9].

The sera of all HIV-infected individuals contains autoantibodies that recognize partly sialylated CD43 molecules ^[10,11]. However, peripheral lymphocytes from HIV-infected individuals are not hyposialylated ^[10], and therefore the origin of the autoimmunogenic CD43 epitope remains unknown. To characterize abnormalities in CD43 expression occurring during HIV infection further, we studied a series of 86 HIV-infected individuals at different stages of disease progression. We found elevated percentages of cells expressing CD43-hexasaccharides on peripheral T lymphocytes of 96.5% of patients, finding CD43 proteins with altered electrophoretic mobility in lymphocyte lysates from all patients analysed.

Materials and methods

Patients

Patients were classified in groups according to their CD4 T cell counts following the criteria of the Centers for Disease Control and Prevention (group 1, CD4 cells/mm³ > 500; group 2, CD4 cells/mm³ 499-200; group 3, CD4 cells/mm³ < 200).

Quantification of T lymphocyte subpopulations

CD3, CD4 and CD8 cell counts were determined in a FacsCount flow cytometer (Becton Dickinson, San Jose, CA, USA) on whole blood samples after double immunofluorescence staining.

Antibodies and flow cytometric analysis

Purified monoclonal antibody (mAb) T305 (which specifically recognizes the CD43-hexasaccharide isoform) and L10 (which reacts with both isoforms) were provided by Dr E. Remold-O'Donnell (Boston, MA, USA). Peridin chlorophyll protein (Per-CP), fluorescein isothianate (FITC) or phycoerythrin-labelled (PE) mAb Leu4 (anti-CD3), L243 (anti-human leukocyte antigen-DR locus molecules; HLA-DR) and L78 (anti-CD69) were obtained from Becton Dickinson. Double or triple immunofluorescence stained samples were analysed after the selective gating of cells in a FACSort Flow cytometer (Becton Dickinson).

Sodium dodecylsulphate-polyacrylamide gel electrophoresis and Western immunoblotting

Peripheral blood lymphocytes separated by centrifugation on Ficoll gradient were lysed with Brij-97 lysing buffer, resolved by sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) (7.5% polyacrylamide gels, reducing conditions) and electrotransferred to Hybond-P hydrophobic polyvinylidene difluoride (PVDF) membranes (Amersham, Buckinghamshire, England). CD43 was probed on PVDF membranes with 0.1 μg/ml of mAb T305 or L10 followed by incubation with horseradish peroxidase-labelled goat anti-mouse antibody second antibody (Transduction Laboratories, Lexington, KY, USA). The blot was developed by chemiluminescence (ECL, Amersham).

Results

We found elevated percentages of T lymphocytes expressing the CD43-hexasaccharide isoform (T305+) in 82 out of 86 HIV-positive individuals (96.5%). CD43-hexasaccharide expression augments as the disease progresses (defined by the decreasing number of CD4 cells) (Table 1), showing statistical significance between every patient group and the control group and between all groups of patients with each other.

Table 1

As the expression of CD43-hexasaccharide is associated with the cell activation state of T lymphocytes ^[12], we examined whether HIV-infected peripheral T lymphocytes T305+ also co-express other activation-related cell surface molecules (Table 1). The identification of cells expressing HLA-DR, carried out by triple immunofluorescence flow cytometric analysis, showed that most HLA-DR-positive T lymphocytes co-express CD43-hexasaccharide on their surface (Fig. 1a). By contrast, a distinct cellular subset is comprised by the majority of T305+ cells, which lack HLA-DR molecules on their cell membrane.

Fig. 1

To define the expression of T305 on peripheral T lymphocyte subpopulations further, we carried out triple immunofluorescence flow cytometric analyses with the selective acquisition of CD3 cells. We found, in nine out of nine patients, that CD43-hexasaccharide expression was markedly skewed towards the CD8 compartment. The T305 epitope was expressed on 24 ± 8% of CD4 T lymphocytes from HIV-infected individuals and on 50 ± 9% of CD8 cells.

We explored the integrity of the CD43 molecule expressed on the surface of HIV-infected individuals because protein instability and the appearance of truncated forms of CD43 are common features of some primary immunodeficiencies ^[13]. A band of 135 000 M_r was detected by Western blot with the T305 mAb in lysates from all HIV-infected individuals (Fig. 1b). This band is not observed with lymphocytes from a healthy individual under adequate times of film exposure (Fig. 1b). Overexposure of the autoradiography allows the appearance of the 135 000 M_r band in normal cells, thus detecting by Western blot the CD43-hexasaccharide expressed on 9% of lymphocytes of the healthy control subject (quantified by flow cytometry). Strikingly, overexposed films detected multiple CD43 bands with altered electrophoretic mobility in lymphocyte lysates from all HIV-infected individuals (Fig. 1b). Similar results were obtained when the membranes were hybridized with the L10 mAb, which reacts with a sialylation-independent epitope present in both tetra and hexasaccharide CD43 isoforms (not shown).

Discussion

We demonstrate in this work that peripheral T lymphocytes from HIV-infected individuals show augmented expression of the CD43-hexasaccharide isoform. Increased CD43-hexasaccharide expression correlates with progression of the disease, and it is biased towards the CD8 compartment.

AIDS patients produce autoantibodies against hyposialylated forms of CD43 ^[11,14], and therefore it was of interest to study cell glycosylation abnormalities caused by HIV. Previous studies focused on T cell lines infected in vitro with HIV, leaving unexplored peripheral T lymphocytes from HIV-positive individuals. We found a 135 000 M_r single band reacting with the T305 mAb, indicating the predominant presence of hexasaccharides on the CD43 expressed on peripheral T lymphocytes from HIV-infected individuals. By contrast, CD43 from in-vitro infected T cell lines migrates on SDS-PAGE as a broad band spanning 110 000-190 000 M_r^[11], indicating that complex, heterogeneous sugars are attached to the protein core. The interference thus caused by HIV to the glycosylation cell machinery leads to the aberrant expression of CD43.

Increasing experimental evidence suggests that CD43 isoforms play differential functional roles. Natural killer cell subpopulations with distinct functional properties selectively express one or the other CD43 isoform ^[5]. Resting T lymphocytes switch upon specific activation the 115 000 M_r CD43 isoform into the 130 000 M_r molecule ^[12], and the disregulated expression of CD43 leads to immunodeficiency ^[15]. In this context, the CD43-hexasaccharide molecules expressed by peripheral lymphocytes of HIV-positive individuals could be involved in the HIV-reactive specific immune response(s). This is supported by our finding of two distinct subpopulations expressing the T305 epitope. The first subset co-expresses HLA-DR and T305, and it might represent activated cells involved in ongoing cytotoxic responses. The second subpopulation, quantitatively predominant, solely expresses T305, and it could reflect lymphocytes that have acquired the molecule to contribute to the regulation of the immune response. Evidence for the regulatory role of CD43 glycans comes from transgenic mice overexpressing core 2-O glycans, which show a reduction in the intensity of primary T cell immune reactions, and it is therefore believed that the surface expression of these molecules would avoid uncontrolled cellular responses ^[16]. In this context, we interpret the fact that the four individuals with the highest CD4 peripheral counts in our series (including two subjects with acute infection and one long-term progressor) have a very high percentage of the surface expression of CD43-hexasaccharide, clearly above the average of CD43-hexasaccharide expression for patients of group 1 (not shown).

As has been mentioned before, all HIV-infected individuals make autoantibodies to hyposialytated forms of CD43. However, the immunogen(s) responsible for this phenomenon remains unidentified because the prime candidate, peripheral lymphocytes of HIV-positive subjects, are not hyposialytated ^[10,14]. It was therefore suggested that autoantigenic CD43 epitopes would be generated during the accelerated process of lymphocyte destruction occurring in the peripheral circulation of HIV-positive individuals ^[10]. Other authors, using HIV-positive T cell lines, proposed that those autoantibodies would recognize a conformational structure on CD43 exposed by the glycosylation changes induced by the viral infection. Finally, alveolar macrophages from HIV-positive individuals bind peanut agglutinin lectin slightly better than their cellular counterparts from normal individuals, indicating a diminished presence of sialic acids on the membrane of this cell subpopulation ^[17]. We have detected in all HIV-positive individuals tested the appearance of bands reacting with both the T305 mAb and the L10 mAb (not shown), which show altered electrophoretic mobility. If these bands are modified CD43 molecules in HIV-positive individuals, they could represent CD43 proteins exposing cryptic epitopes that are usually masked by the heavy O-glycosylation of the molecule, thus preventing autorreactivity. Therefore, our findings favour the hypothesis that the autoimmunogenic determinant is a de novo conformational epitope(s) generated within the altered CD43 glycoproteins. Alternatively, different migration patterns observed may represent full-length proteins with different levels of glycosylation. Nevertheless, the molecular basis for the observed CD43 abnormalities remain to be clarified, along with its precise participation in the pathogenic mechanisms of HIV infection.

Acknowledgements

The authors are indebted to Dr Eileen Remold-O'Donnell for support and the generous gift of reagents.