The protective role of beta-chemokines in HIV infection and disease remains controversial. Contradictory findings have been reported possibly as the result of different beta-chemokine detection methods. To test this, peripheral blood lymphocytes from treatment-naive HIV patients, patients on highly active antiretroviral therapy (HAART), and uninfected controls were assessed for intracellular beta-chemokine levels in comparison with levels of beta-chemokine secretion in culture supernatants. HIV patients had significantly higher intracellular levels of macrophages inflammatory protein (MIP)-1 alpha and MIP-1 beta than uninfected control subjects. In contrast, MIP-1 alpha and MIP-1 beta supernatant levels were significantly lower in HIV patients than in controls. Interestingly, both intracellular and supernatant levels of RANTES (regulated on activation, normal T cell expressed and secreted) were significantly increased in HIV patients. Prolonged (> 3 years) administration of HAART in HIV patients normalized the intracellular levels of MIP-1 beta and RANTES and restored the decreased supernatant levels of MIP-1 alpha and MIP-1 beta to levels observed among controls. Significant direct correlations observed between the intracellular and the supernatant levels of beta-chemokines in controls were lost in treatment-naive (except MIP-1 beta) and HAART-treated patients (except RANTES after 3 years of HAART). These data indicate that lymphocytes of HIV patients display a disrupted capacity to secrete the beta-chemokines MIP-1 alpha and MIP-1 beta, which may constitute a mechanism of immune dysfunction in progressive HIV infection. Furthermore, we demonstrated that the detection of beta-chemokines in HIV patients by different methods may indeed result in contradictory findings.