α1Proteinase inhibitor regulates CD4+ lymphocyte levels and is rate limiting in HIV-1 disease

Abstract

Background: The regulation of adult stem cell migration through human hematopoietic tissue involves the chemokine CXCL12 (SDF-1) and its receptor CXCR4 (CD184). In addition, human leukocyte elastase (HLE) plays a key role. When HLE is located on the cell surface (HLE(CS)), it acts not as a proteinase, but as a receptor for α(1)proteinase inhibitor (α(1)PI, α(1)antitrypsin, SerpinA1). Binding of α(1)PI to HLE(CS) forms a motogenic complex. We previously demonstrated that α(1)PI deficiency attends HIV-1 disease and that α(1)PI augmentation produces increased numbers of immunocompetent circulating CD4(+) lymphocytes. Herein we investigated the mechanism underlying the α(1)PI deficiency that attends HIV-1 infection.

Methods and findings: Active α(1)PI in HIV-1 subjects (median 17 µM, n = 35) was significantly below normal (median 36 µM, p<0.001, n = 30). In HIV-1 uninfected subjects, CD4(+) lymphocytes were correlated with the combined factors α(1)PI, HLE(CS) (+) lymphocytes, and CXCR4(+) lymphocytes (r(2) = 0.91, p<0.001, n = 30), but not CXCL12. In contrast, in HIV-1 subjects with >220 CD4 cells/µl, CD4(+) lymphocytes were correlated solely with active α(1)PI (r(2) = 0.93, p<0.0001, n = 26). The monoclonal anti-HIV-1 gp120 antibody 3F5 present in HIV-1 patient blood is shown to bind and inactivate human α(1)PI. Chimpanzee α(1)PI differs from human α(1)PI by a single amino acid within the 3F5-binding epitope. Unlike human α(1)PI, chimpanzee α(1)PI did not bind 3F5 or become depleted following HIV-1 challenge, consistent with the normal CD4(+) lymphocyte levels and benign syndrome of HIV-1 infected chimpanzees. The presence of IgG-α(1)PI immune complexes correlated with decreased CD4(+) lymphocytes in HIV-1 subjects.

Conclusions: This report identifies an autoimmune component of HIV-1 disease that can be overcome therapeutically. Importantly, results identify an achievable vaccine modification with the novel objective to protect against AIDS as opposed to the current objective to protect against HIV-1 infection.

Conflict of interest statement

Competing Interests: Co-author RW is President of the Institute for Human Genetics and Biochemistry which oversees the Harry Winston Research Foundation that funds research conducted by lead author CB. CB acts as a research consultant to the Institute. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials. All other authors declare no competing interests.

Figures

Figure 1. Correlation between α1PI, IgG-α1PI immune complexes, and CD4+ lymphocytes in HIV-1 infected subjects.

(A) In subjects with >220 CD4 cells/µl, CD4+ lymphocyte levels correlate with active α1PI (r2 = 0.927, p<0.0001, n = 26). CD4+ lymphocyte levels also correlate with inactive α1PI, (r2 = 0.906, p<0.0001, n = 26). Subjects receiving protease inhibitor therapy are depicted by red squares. All other subjects are depicted by black circles. In the 9 subjects with <220 CD4 cells/µl, no correlation was found to exist between CD4+ lymphocyte levels and active α1PI. Non-linear regression was performed using a 3 parameter Sigmoid curve with power of test α = 0.05. In this population, all variables were found to have normality and constant variation. (B) In 8 of 35 subjects, IgG-α1PI immune complexes were detected and were correlated with CD4+ lymphocyte levels (r2 = 0.822, p = 0.05) and with inactive α1PI (r2 = 0.988, p<0.0001).

Figure 2. Binding of anti-gp120 to human, but not chimpanzee α1PI.

(A) Monoclonal antibody 3F5 (5 µg/ml) binding to α1PI in sera from 18 HIV-1 uninfected humans and 20 HIV-1 uninfected chimpanzees was measured using ELISA. Antibody bound (A490 nm) was normalized for the active α1PI concentration in each specimen and is represented as A490nm/[α1PI (µM)]. Representative data from 6 measurements are depicted. Bars represent median values. Median 3F5 bound to human α1PI was 0.12 and to chimpanzee α1PI was 0.02. Negative control monoclonal antibody α70 (10 µg/ml) yielded A490nm = 0.02 when incubated with α1PI at concentrations varying between 3 µM and 540 µM. There was no difference in binding of α70 to human or chimpanzee sera (p>0.6). (B) IgG-α1PI immune complexes (A490 nm) were measured in sera from HIV-1 uninfected humans (n = 9), HIV-1 infected humans (n = 35), HIV-1 uninfected chimpanzees (n = 20), HIV-1 challenged chimpanzees (n = 2), rhesus monkeys pre-immunization and 2 time points post immunization (n = 12), and rhesus monkeys pre- and post-infection (n = 3). There was no significant difference in rhesus monkeys pre- and post-immunization, pre-and post-infection. Representative data of triplicate measurements are depicted. (C) Active α1PI was measured in HIV-1 uninfected humans (26 µM, n = 20), HIV-1 infected humans (18 µM, n = 35), HIV-1 uninfected chimpanzees (35 µM, n = 20), HIV-1 challenged chimpanzees (39 µM, n = 2), rhesus monkeys pre-immunization and 2 time points post immunization (36 µM, n = 12), and rhesus monkeys pre- and post-infection (43 µM, n = 3). There was no significant difference in rhesus monkeys pre- and post-immunization, pre-and post-infection. Bars represent median values. (D) Inactive α1PI was measured in HIV-1 uninfected (4 µM, n = 20) and HIV-1 infected humans (19 µM, n = 35). Bars represent median values. (E) Active α1PI levels in sera from 5 HIV-1 infected subjects after incubation with either medium (control) or with monoclonal antibody 3F5.

Figure 3. Corresponding conformation at the 3F5-recognized epitope in α1PI and CD4-complexed HIV-1 gp120.

HIV-1 gp120 is depicted from two perspectives (A,B) with green representing two α-helices (aa 100–115 and 476–484). The gp120 peptide immunogen used to raise 1C1 and 3F5 (aa 471–490) is located at the C-terminus of gp120, and the linear segment 486YKVV489 is depicted in red along with Met95 and the oligosaccharide-linked segment 234NGT236, all of which are within 8 Å of the conformational epitope. The gp120-homologous domain in α1PI is also located at the C-terminus of the protein, and is depicted from two perspectives (C,D) with violet representing the antiparallel β-sheet strand at the base of the cleft (aa 369–389), and green representing the α-helices that form the mouth of the cleft (aa 27–44 and 259–277). Met-385, which distinguishes human from chimpanzee α1PI, is depicted in red along with the segment 386GKVV389, the oligosaccharide, and oligosaccharide-linked segment 46NST48. The HLEG-reactive site Met-358, is depicted in yellow for orientation. Structures for human α1PI (1HP7) and CD4-complexed HIV-1 gp120 (1RZJ) from the NCBI Molecular Modeling Database (MMDB) were analyzed using Cn3D software. Small carbohydrate structures, depicted in multiple colors, were associated with 1RZJ in MMDB, and the three associated with 1HP7 were added using Adobe Photoshop.