This study of ART-negative HIV patients demonstrated dramatically increased production of 9G4+ IgG serum antibodies () and the in vivo autoreactivity of these 9G4 antibodies as indicated by their binding to B cells (+B). This increase was positively correlated with HIV VL (). There was also a significant expansion of 9G4+ memory B cells that had distinct phenotypes compared with 9G4− memory B cells as indicated by IgG+ and IgM+ memory B cell populations being over-represented in the 9G4+ memory compartments (+D). A fraction of the 9G4+ serum antibodies and B cells were reactive to HIV Env (+4), and the abundance of 9G4+ serum antibodies and B cells positively correlated with serum HIV neutralizing activity (). These findings demonstrate the association of autoreactivity with the development of HIV BNAb at both the cellular level and at the level of serum antibodies.
The presence of 9G4+ IgG serum antibody and 9G4+ memory B cells is consistent with the impaired B cell tolerance in HIV patients which has been previously observed, primarily as autoreactive serum antibodies 
. Although the presence of elevated 9G4+ IgG serum appears to be slightly greater (40%) as compared to anti-dsDNA, ANA, and CL (15%, 24%, 31% respectively) in our cohort. Furthermore, we observed a significant positive correlation between HIV VL and 9G4+ IgG antibody and 9G4+ B cells, but no significant correlation between VL and dsDNA, ANA, or CL reactive IgG antibodies. This suggests that 9G4 induction may be a more selective process induced by the HIV virus, either directly through viral antigens or indirectly through the generation of autoantigens known to react with 9G4+ antibodies, such as apoptotic cells 
, as opposed to simply generalized B cell activation which might, at least in part, lead to the development of other autoreactivities in HIV patients, such as anti-DNA and anti-CL antibodies. The former possibility is being currently addressed by the characterization of 9G4 monoclonal antibodies generated from HIV patients using single-cell methodologies, and will additionally enable the assessment of polyreactivity within the 9G4 compartment. The role of the virus in the induction of the 9G4 responses could also be clarified in part by ongoing longitudinal analysis of HIV patients, before and after the initiation of ART, or starting during acute infection to examine the initial emergence of the 9G4 response, however, this is beyond the scope of the current study.
Although 9G4+ B cells shared many phenotypic commonalities with 9G4− B cells in HIV patients, including their presence in both naïve and memory B cell compartments, they were over-represented in the “IgM only" memory (IgD−IgG−IgM+CD27+) compartment compared to 9G4− B cells (). The rare IgM only memory population, is expanded in patients deficient in activation-induced cytidine deaminase (AID) 
and has been suggested to result from cells exiting prematurely from the GC reaction before class switching occurs or may be due to the result of GC-independent B cell development 
. This would be consistent with the features of 9G4+ B cell regulation, where in healthy control subjects, they are fully excluded from the GC. However in HIV patients, some 9G4+ B cells are likely to have participated fully in GC reactions, as evidenced by IgD−IgG+ 9G4+ B cell populations. Other 9G4+ B cells in HIV patients may still be subject to intrinsic or extrinsic regulation that prevents full GC participation, as evidenced by the over-representation of the IgM only memory population within the 9G4+ B cell compartment. It is tantalizing to speculate on properties that may distinguish 9G4+ IgM only memory vs. 9G4+ IgD−IgM−IgG+ memory B cells, such as degree of autoreactivity, responsiveness to T cell help or suppression, or capacity to migrate toward CXCL13. These possibilities will be dissected in future studies. Additional phenotypic differences in the 9G4+ B cell population may exist and may be revealed by more comprehensive flow-cytometric based profiling.
The observation of a higher frequency of 9G4+ cells in the Env-specific memory B cell population compared with the total memory population suggests in some HIV patients antigen-specific 9G4+ expansion may be occurring. This observation warrants further investigation to determine if such patients have unique features, and to address the possible mechanism of 9G4+ Env-specific memory B cell expansion. The positive correlation observed between serum HIV neutralizing activity and 9G4+ IgG may reflect a direct contribution of 9G4+ IgG to HIV neutralizing activity, or it might reflect an immunological environment, in which HIV infection coincidently favors the expansion of both species. It is evident that 9G4+ cells and serum antibodies represent only a minority of the Env-specific repertoire, but it remains to be determined if 9G4+ Env-specific cells and antibodies have any distinct properties beyond inherent autoreactivity that may uniquely impact HIV. Although numerous HIV broadly neutralizing monoclonal antibodies have been described to date, none are VH4-34-encoded. This does not preclude their existence or the potential of 9G4+ B cells and antibodies to contribute substantially to a protective immune response to HIV. Indeed, given the reactivity 9G4+ antibodies exhibit for glycoproteins, including an N-linked N acetyllactosamine determinant on CD45R/B2220 
and the I/I blood group antigen 
, reactivity of 9G4+ antibodies to HIV Env may in part be mediated through interaction with glycosylated epitopes. Thus, detailed assessment of the 9G4+ serum antibodies for HIV neutralizing activity and specificity should be pursued in future studies. The expansion of 9G4 in HIV patients could also result from the polyclonal B cell activation that is readily observed in HIV patients, which may result from numerous factors including direct interaction of HIV and B cells through CD21, DC-SIGN, TLR7, and TLR9 and also indirectly as a consequence of cytokine upregulation (e.g. IL4, IL10, IL6, IFNα) and T cell help 
This work demonstrates the utility of the 9G4 system in interrogating autoreactivity at the cellular level in HIV patients, enabling insight into the development and expansion of autoreactive B cells and their relationship to the serum antibody repertoire. Important future questions include whether autoreactive antibodies are a prerequisite for the development of a protective HIV humoral response. There are clear examples of individual BNAbs against HIV that have minimal autoreactivity (e.g. VRC01) 
, but it is unclear if there is a role for relaxed tolerance in promoting their development. And although it is likely that autoreactivity in HIV patients may be in part a consequence of B cell polyclonal activation, polyclonal activation may also greatly expand the available repertoire of HIV-specific B cells. For example, a hyperactive state of global B cell dysregulation and functional threshold decreases may allow B cells that bind only weakly to HIV-1 Env to productively respond to this viral antigen, and give rise to antibody-secreting cells.
The association of autoreactivity and BNAbs also has important implications for HIV vaccine development – especially if autoreactivity is indeed a prerequisite for the development of a subset of immunoglobulin specificities that have the potential to give rise to BNAbs. Although it may not be feasible to completely uncouple autoreactivity from BNAb development in a vaccine setting, it is important to note that autoreactivity does not necessarily equate to manifestations of chronic autoimmune pathology, or a definitive pre-disposition to autoimmune disease development. Several observations support this, including the development of “non-pathological" autoreactive antibodies during various viral infections including CMV, HCV, and RSV 
. Consistent with this, a number of investigators have described autoreactive antibodies in HIV-infected individuals as ‘non-pathogenic", based on distinct reactivity profiles 
or lack of β2GP1 involvement 
- as compared with HIV-negative patients with primary autoimmune diseases. Indeed, autoreactive antibodies are frequently observed in otherwise normal healthy individuals, absent of any clinical autoimmune disease manifestations. Minimally, the association of auto- and HIV- reactivity warrants detailed examination, which can be facilitated by the 9G4 system, as it may reveal critical cellular and molecular mechanisms by which the repertoire of potential BNAb specificities is effectively engaged and stimulated to develop into long-lived plasma cells conferring sustained protection from HIV infection.