The pathogenesis of bone metabolism disorders in HIV-infected patients has been attributed to either HIV or several antiretroviral drugs, such as tenofovir or protease inhibitors [4–6, 25]. Recently, a pathogenetic model attributed the early onset of osteopenia/osteoporosis to the generalized inflammation and immunosenescence that characterizes HIV-infected populations, including in the context of virologically suppressive HAART [12, 13].
In this study, we investigated whether HIV-infected patients who were affected by osteopenia/osteoporosis displayed a circulating T-cell immune phenotype that indicated hyperactivation and premature senescence.
In HIV-infected patients who were affected by low BMD, we demonstrated the following: 1) higher peripheral T-cell activation, 2) less pro apoptotic CD28-negative CD4+ T-cells despite no changes in senescent CD28-57+ T-cells, and 3) a highly activated T-cell phenotype as an independent predictor of impaired BMD.
Elevated immune activation is the main pathogenetic mechanism that is involved in HIV infection with and without HAART: it is predictive of the clinical progression to AIDS  and has recently been demonstrated to be associated with the emergence of co-morbidities, such as cardiovascular disease and intima thickness [26, 27].
For the first time, our study demonstrates the involvement of HIV-related immune activation in the pathogenesis of bone co-morbidities. The immune phenotype in the peripheral blood of patients with pathological BMD is characterized by a higher frequency of activated T-cells compared with patients with normal BMD. In addition, the association between T-cell activation and low BMD was confirmed in patients on stable HAART with full virological suppression. Interestingly, in the multivariate logistic regression analysis that was adjusted for demographics and HIV- and HAART-related parameters, a higher frequency of activated CD4+/CD8+ T-cells was an independent predictor of osteopenia/osteoporosis.
Similarly, we observed that the immune phenotype in the peripheral blood of patients with impaired BMD was characterized by a non significant trend toward lower CD127-expressing central memory CD8+ T-cells compared with patients with normal BMD.
In HIV/AIDS, the expansion of CD8+ T-cells that lack CD127 has been consistently described [28–31] as a functional status of terminal differentiation [32, 33] and has been attributed to chronic antigenic stimulation.
Thus, higher T-cell activation with contraction in the compartment of central memory CD127 + CD8+ T-cells in patients who are affected by osteopenia/osteoporosis may suggest increased T-cell turnover with skewed maturation of CD8+ T-cells.
Osteoporosis is not typically considered to be an immune-mediated disorder; however, recent data have indicated an overlapping pathway between bone physiology and the biology of inflammation that involves the T-lymphocyte compartment. Activated T-cells affect bone physiology by producing RANKL and pro-inflammatory cytokines (e.g., IL-1 and TNF-α), which promote osteoclast activity  and stimulate stromal cells to produce osteoclastogenic IL-7 .
Thus, after observing a hyperactivated circulating T-cell immune profile in patients with bone metabolism disorders, we investigated the pro-inflammatory osteoclastogenic cytokine milieu.
Despite a non significant trend toward a higher level of TNF-α in patients with low BMD, the levels of IL-7 and osteoclastogenic cytokines were comparable among LBMD and NBMD subjects. More than 70% of the patients were on long-term HAART; therefore, this finding is in agreement with recent data that demonstrated an increase in the osteoclastogenic cytokines RANKL and TNF-α within the first weeks of therapy with no additional surges . Moreover, the DXA analysis provides a static picture of BMD; however, the cytokine milieu depicts the current balance of bone remodeling processes and may be influenced by many factors, mainly the effect of protease inhibitors .
We next determined whether these LBMD patients were characterized by an immune senescent pattern in peripheral blood.
Despite increased activation, no differences in the proportion of senescent (CD28-CD57+) CD4+ and CD8+ T-cells were observed in patients with impaired BMD compared with patients with normal BMD.
Unexpectedly, patients with impaired BMD displayed a circulating CD4+ T-cell immune phenotype that was enriched for CD28-expressing cells at the disadvantage of the highly differentiated CD28-negative pool. Interestingly, this CD4+ immune phenotype was independently associated with osteopenia/osteoporosis in the multivariate logistic regression analysis.
The downregulation of CD28 has been extensively described in CD8+ T-cells as a hallmark of chronic antigenic stimulation, even in the course of HIV infection. With each repetitive stimulation/proliferation round, CD28 expression is irreversibly downregulated on the CD8+ surface, leading to the accumulation of antigen-experienced (CD28-negative) CD8+ T-cells with shortened telomeres, impaired proliferative ability and decreased susceptibility to apoptosis.
In addition, the loss of CD28 is observed in CD4+ T-cells during chronic immune activation but at a significantly lower rate than CD8+ . Despite the evidence of CD28-CD8+ T-cell accumulation in HIV infection, there are no data on CD28-CD4+ T-cell accumulation.
In our study, patients who were affected by osteopenia/osteoporosis displayed high levels of T-cell activation and a reduction in the CD28-CD4+ T-cell pool. As a possible explanation for this finding, Vivar et al. recently demonstrated that CD28-negative T cells, although senescent, display a proapoptotic phenotype, as suggested by high Fas and decreased IL7Rα and Bcl-2 expression . These T-cells are highly susceptible to apoptosis, which results in increased apoptosis during HIV replication and generalized immune activation. Therefore, the contraction of the CD28-negative CD4+ T-cell pool that we observed in patients with bone disorders may have resulted from increased activation-induced apoptosis.
The main limitation to this study is the population that was enrolled, which included both HIV-infected patients who were naïve to antiretroviral therapy and patients undergoing HAART. Many factors may have played a role in the pathogenesis of osteoporosis in these two distinct populations because several antiretroviral medications may directly decrease bone mineral density and act as confounding factors.