Virological and Immunological Outcomes of Combined Therapeutic Interventions and Dendritic Cell Therapy in People With HIV

Abstract

Background

Except for hematopoietic cell transplantation, therapeutic-driven HIV-1 curative approaches have been unsuccessful. Here, we describe a 2-step randomized clinical trial designed to evaluate the safety and impact of individual and combinatorial therapeutic strategies on changes in peripheral and gut mucosal HIV reservoirs, immune activation and function in people with HIV with chronic disease and high CD4 T-cell nadirs.

Methods

Thirty participants were enrolled and randomized equally into 6 study arms based on treatments with either standard antiretroviral therapy (ART) alone or a combination of candidate anti-HIV reservoir strategies that included ART intensification, auranofin (an apoptotic-inducer antirheumatic drug), nicotinamide (vitamin B3), and a personalized dendritic cell therapy.

Results

After an analytical treatment interruption post intervention, all eligible participants rebounded within 14 weeks, except for 1 participant with rebound detected at 84 weeks and 2 participants receiving all combined therapies maintained viral loads below 1000 copies/mL through the study period. These posttreatment controllers all received nicotinamide containing regimens and exhibited immune cellular epigenetic age reversal.

Conclusions

This proof-of-concept clinical trial demonstrates the safety of multiple interventions with distinct antireservoir activities in people with HIV and argues for continued investigation of both single and combinatorial interventions towards posttherapy HIV control.

This proof-of-concept clinical trial using a combination of therapeutic interventions randomized among antiretrovirally suppressed individuals with HIV was clinically safe and, following an analytical treatment interruption, led to a reduction in viral load set point and several cases of posttreatment HIV control.

Despite anecdotal reports, no long-term antiretroviral therapy (ART)–free HIV remission strategy can be readily employed in clinical practice. Barriers to eliminating HIV in people with HIV (PWH) on suppressive ART include residual viral replication, viral sanctuaries, and cells harboring transcriptionally silent but replication-competent-integrated HIV DNA, considered HIV reservoirs [1]. Any attempt toward HIV clearance in the absence of ART will involve eliminating or controlling latent HIV reservoirs. In addition, it will be necessary to mitigate the residual micro-inflammation associated with HIV protein production from defective proviruses in PWH despite suppressive ART [2]. Accordingly, multiple strategies, in addition to ART, are likely necessary to control, decrease the size of, and eventually eliminate the HIV reservoir.

There several lines of evidence supporting the safety and feasibility of HIV Gag–targeted immunologic strategies. The randomized trial ACTG A5197 tested an adenovirus-5 (Ad5)–based HIV Gag vaccine and reported a correlation between post-ATI viral load set point and the extent of anti-HIV Gag cellular immune responses [3]. Moreover, an early attempt of personalized cell therapy exploiting, among others, anti-HIV Gag immune responses was conducted using autologous dendritic cells (DCs) transfected with RNA encoding the HIV-1 Gag, Vpr, Vif, and Nef proteins [4]. However, though immunogenic, this cell therapy did not result in post-ATI viral load containment, suggesting that combined interventions are required in addition to personalized cell therapy.

In the present study, we describe the impact of selected, isolated, and combined strategies on viral persistence and systemic inflammation in PWH who had been receiving suppressive ART regimens for over 2 years without advanced HIV disease and displaying immune reconstitution of CD4+ T-cell count (nadir and current CD4 > 350 cells/mm³). Based on preclinical and clinical data, the strategies employed included an ART intensification regimen with dolutegravir with and without maraviroc to further intensify antiretroviral activity and to attempt to disrupt HIV latency [5]; use of the gold salt auranofin to induce apoptosis selectively in memory T cells [6, 7]; use of nicotinamide to potentially revert HIV latency [8] and decrease microinflammation [9]; and use of an autologous, personalized, HIV monocyte–derived DC therapy pulsed with autologous HIV Gag peptides to boost specific cellular immunity [10]. Notably, in this trial, we restricted the immune stimulus to highly conserved regions of autologous HIV-1 Gag and to the peptides predicted to bind each individual's class I HLA. This choice was based on previous evidence from studies on human HIV elite controllers and nonhuman primates, that cell-mediated immunity against Gag, and not other viral antigens, was the key correlate of reduced plasma viral loads in the absence of ART [11].

METHODS

Study Design and Participants

Participants were recruited into a single site in São Paulo, Brazil. The Federal University of São Paulo institutional review board (IRB) approved the trial, and participants provided informed written consent (UNIFESP IRB #10757312.0.0000.5505). The trial was registered with ClinicalTrials.gov (NCT02961829). Auranofin (Astellas Pharma, Milan, Italy) was prescribed at 3 mg BID and nicotinamide at 1 g BID. Participants who received DC therapy were not prescribed maraviroc due to its likely effects on the chemotaxis of DCs [12]. An interim analysis evaluating the impact of auranofin on the HIV reservoir has been previously reported [7]. Dendritic cell therapy was prepared as previously described [10]. Participants were followed every 4 weeks during the intervention period. Inclusion criteria included age > 18 years, stable first-line ART (changes for toxicity allowed) for at least 2 years, use of the same ART regimen for at least 1 year, CD4+ T-cell count >500 cells/mm³, nadir CD4+ T-cell count >350 cells/mm³, plasma HIV RNA below the limit of detection since ART suppression, and presence of CCR5-tropic HIV-1 at screening as inferred by proviral DNA genotropism (see below). Exclusion criteria are outlined in Figure 1. The primary endpoints of this trial were to determine changes in VL, cell-associated HIV RNA, episomal DNA, HIV-specific antibody titers, and CD4 and CD8 T-cell activation.

Figure 1.

Participant flow diagram and trial design. Flow diagram of participant inclusion and randomization (A). SPARC-7 trial design and analysis (B). ART, antiretroviral treatment; ATI, analytical treatment interruption. Screening commenced on 24 August 2015, and the last study follow-up visit for the 48-week intervention period occurred on 28 August 2017. Personalized DC therapy for participants in arms E and F commenced on 14 August 2017, and the last participant received DC therapy on 8 December 2017. Subsequently, study participants were maintained on the same ART regimen they received before study initiation, with follow-up visits every 3 months.

Clinical, Laboratory, and Statistical Procedures

Please, see Supplementary Data.

RESULTS

Of the 64 screened potential participants, 10 harbored non-CCR5-tropic viruses, 6 harbored viruses with indeterminate genotropisms, 6 withdrew consent before randomization, and 12 failed screening for various clinical reasons. The remaining 30 candidates fulfilled all inclusion and exclusion criteria and were randomized to the 6 study arms each with 5 participants, using a computer-generated list (Figure 1 and Table 1; Supplementary Methods). The control group (arm A) continued the same ART regimen. The ART intensification group (arm B) received dolutegravir and maraviroc added to their ART regimen for 48 weeks. The ART intensification + nicotinamide group (arm C) received antiretroviral intensification (maraviroc and dolutegravir) and nicotinamide for 48 weeks. The ART intensification + auranofin group (arm D) received auranofin for the first 24 weeks of the study and ART intensification (maraviroc and dolutegravir) for 48 weeks. The dolutegravir + DC therapy group (arm E) received a partially intensified ART regimen for 48 weeks (dolutegravir only), followed by DC therapy (see Methods, Supplementary Figure 1) administered in 3 doses: at the end of 48 weeks and 2 and 4 weeks thereafter. Finally, the dolutegravir + nicotinamide + auranofin + DC therapy group (arm F) received auranofin for the first 24 weeks, including both nicotinamide and a partially intensified ART with dolutegravir for 48 weeks, followed by DC therapy administered in 3 doses as outlined in arm E above.

Table 1.

Study Participant Characteristics: Multi-Interventional Trial

Study Arm		PID	Age (y)	Race	Sex	Pre-ART HIV-1 RNA Level Log10 (Copies/mL)	HIV Clade (V3)a	ARV Regimen at Study Enrollment	ART Duration at Study Enrollment (y)	Baseline CD4+ T-Cell Counts (Cells per Mm3)	Postintervention CD4+ T-Cell Counts (Week 48)	CD4+ T-Cell Nadir	Notes
ART		1	30	Latin-White	M	4.90	B	TDF/3TC/EFV	5	541	636	395
		2	49	Latin-White	M	4.24	B	TDF/3TC + FPV-r	15	788	791	534
	(A)	3	34	Latin-White	M	5.04	C	TDF/3TC + NVP	5	1067	1288	480
		4	54	Latin-White	M	4.72	B	TDF/3TC/EFV	5	775	743	658
		5	34	Latin-White	M	3.99	B	TDF/3TC + RAL	3	1585	1482	758
ART intensification		6	29	Latin-White	M	4.07	C	TDF/3TC/EFV	5	1287	716	492
		7	61	Latin-White	M	4.99	B	AZT/3TC + NVP	11	1202	972	566
	(B)	8	40	Latin-White	M	3.88	B	AZT/3TC + ATV-r	5	921	861	604
		9	36	Latin-White	M	4.74	B	TDF/3TC/EFV	6	805	1048	661
		10	45	Latin-White	M	4.14	B	TDF/3TC/EFV	5	950	1166	683
ART intensification + nicotinamide		11	40	Latin-White	M	4.27	B	TDF/3TC/EFV	6	762	780	398
		12	28	Latin-White	M	3.52	B	ABV + 3TC + ATV-r	6	1231	1018	731
	(C)	13	34	Latin-White	M	4.31	B	TDF/3TC/EFV	6	720	698	372
		14	52	Latin-White	M	4.95	A	AZT/3TC + EFV	8	964	914	490
		15	35	Latin-White	M	3.73	B	TDF/3TC/EFV	5	646	934	506
ART intensification + auranofin		16	41	Latin-White	M	4.20	B	AZT/3TC + EFV	8	1274	1467	412
		17	32	Latin-White	M	4.31	B	ABC + 3TC + EFV	6	625	726	372
	(D)	18	48	Latin-White	M	4.35	B	AZT/3TC + EFV	11	1335	1313	873
		19	35	Latin-White	M	4.94	B	TDF/3TC + FPV-r	6	1016	967	434
		20	28	Latin-White	M	5.19	B	TDF/3TC + ATV-r	6	979	984	448
ART partial intensification + DC therapy		21	32	Latin-White	M	4.99	B	TDF/3TC + FPV-r	10	1053	1061	749
		22	55	Latin-White	M	5.56	B	AZT + 3TC + EFV	7	1404	1361	552
	(E)	23	27	Latin-White	M	5.72	B	TDF/3TC + ATV-r	5	693	871	490
		24	56	Latin-White	M	5.68	B	TDF/3TC/EFV	22	874	879	161	Protocol violator
		25	44	Latin-White	TGW	5.01	B	TDF/3TC + LPV-r	21	1591	1191	400
Multi-interventional		26	38	Latin-White	M	4.11	F	AZT/3TC + EFV	6	770	847	530
		27	25	Latin-White	M	5.24	B	TDF/3TC/EFV	3	839	915	766	Protocol violator
	(F)	28	45	Latin-White	M	4.76	C	TDF/3TC/EFV	6	1265	1097	685
		29	60	Latin-White	M	4.52	B	TDF/3TC/EFV	8	677	720	505
		30	61	Latin-White	M	5.44	B	TDF/3TC/EFV	13	1001	930	460

Abbreviations: TDF, tenofovir; AZT, zidovudine; ABV, abacavir; EFV, efavirenz; FPV-r, ritonavir-boosted fosamprenavir; NVP, nevirapine; ATV-r, ritonavir-boosted atazanavir; LPV-r, ritonavir-boosted lopinavir. “/” indicates fixed-dose combination.

^aHIV clade was assigned according to the C2-V3-C3 profile of gp120.

Individual baseline study participants’ demographic and clinical characteristics are detailed in Table 1. All participants were Latin-White males with a mean age of 38.3 years (23–58). The median nadir CD4+ T-cell counts and VL before ART initiation were 506 cells/μL (range 161–873) and 53.656 HIV RNA copies/mL (range 3.300–527.795), respectively. The median CD4+ T-cell count at study entry was 950 cells/μL (range 541–1617), and VL was target not detected (TND) for all study participants. In addition, the mean time on ART was 7.7 years (range 3–22). Participant P24 from study arm E initially submitted documents demonstrating a CD4+ T-cell count nadir > 350 cells/μL. Nonetheless, in a further review of the Brazilian National Database for VL and CD4+ T-cell counts performed after the intervention period, a lower nadir of 161 cells/μL was found. Participants were determined to predominantly have subtype B (83.3%) (Supplementary Figure 2) and, except for participant P17, all were CCR5 wild-type homozygous (Supplementary Table 1). After sequencing HIV-1 Gag (Supplementary Figure 3) and determining HLA profiles (Supplementary Table 2) for each participant, autologous Gag peptides were selected based on the predicted individual immunogenicity and used to pulse DCs.

All participants adhered to the protocol during the 48-week intervention period. During the intervention period (before DC therapy), grade 3 or 4 adverse events were not observed (Supplementary Table 3). Furthermore, no significant changes in CD4+ T-cell counts, CD8+ T-cell counts, or CD4/CD8 T-cell ratios (Figure 2A–C) nor major HIV RNA blips (>150 copies/mL; Figure 2D) or confirmed virologic failures were observed.

Figure 2.

Changes in T-cell counts and viral load during the interventional period. Change in absolute CD4 (A) and CD8 (B) T-cell counts, CD4/CD8 ratio (C), and plasma HIV-1 RNA viral load (D) is shown according to the intervention groups.

We next assessed changes in cell-associated HIV following 48 weeks of intervention in all arms and at the 3rd dose of DC therapy for arms E and F. In these 2 arms, however, 2 participants (P24 and P26; Table 1) deviated from the protocol as they opted to interrupt ART after week 48 and before DC therapy unbeknownst to the investigative team and thus were removed from some of our analyses. We first measured the effects of these interventions on PBMC HIV RNA levels and observed no changes across all arms (intention-to-treat analysis; Figure 3A). Next, we measured the effects of the interventions on HIV episomal DNA and observed that baseline episomal DNA levels were undetectable in several donors across arms, thus precluding the possibility of estimating the impact of the interventions on this parameter (Supplementary Figure 4). We thus proceeded with the measurement and assessment of total cellular HIV DNA, although this was not prespecified as a primary outcome. In participants in arms A–F, no changes were detected in the mean total HIV DNA in PBMC or in rectal mononuclear cells during the intervention period from baseline to week 48 (Figure 3B). In arms E and F following DC therapy, no differences were observed in either total PBMC or rectal HIV DNA levels. However, in arm F, 2 participants (P27, P29) from baseline presented with changes in PBMC HIV DNA to below the limit of detection, and both also displayed low or undetectable levels of rectal HIV DNA post-DC therapy (Figure 3B and 3C).

Figure 3.

Cell-associated HIV RNA and DNA assessments in peripheral and rectal cells during interventional period and post-DC therapy. Cell-associated (CA) RNA (A) and total DNA (B) in PBMCs and in rectal tissue (C). CA-RNA and total DNA in PBMC time point differences per group were evaluated by repeated measures 1-way ANOVA with Geisser-Greenhouse correction and Tukey test for multiple comparisons. Total HIV DNA in rectal tissue differences were evaluated by paired t-test.

Next, we analyzed the effects of the multi-interventions on immunological outcomes. We assessed for changes in T-cell activation by assessing the expression of activation markers CD38 and HLA-DR in CD4+ and CD8+ T cells (intention-to-treat analysis; Supplementary Figure 5A). In arm E, we observed decreased expression of CD38 and HLA-DR on CD4+ T cells between week 0 and week 48 (P = .0342 and P = .0133, respectively; Supplementary Figure 5B). A decrease in HLA-DR expression on CD4+ T cells from week 0 to week 48 was also observed in arm F (P = .0047). A decrease in CD38 expression between weeks 0 and 48 as well as an increase in HLA-DR between weeks 48 and 60 was observed on CD8+ T cells in arm F (P = .0047 and P = .0342, respectively; Supplementary Figure 5C). HLA-DR trended to decreases on both CD4+ and CD8+ T cells in arm C. However, this did not impact the immunogenicity of DC therapy as demonstrated in ELISPOT responses measuring personalized HIV-Gag–specific T cells, for which participants in arm F displayed higher responses when compared to those in arm E: arm E versus arm F (P < .0001, F[1,4] = 725.2, degrees of freedom [df] = 1); unstimulated versus stimulated (P = .0005, F[1,4] = 108.0, df = 1); and (arm E vs arm F) × (unstimulated vs stimulated) (P = .1196, F[1,4] = 143.5, df = 1; Supplementary Figure 6A).

We have previously shown that quantitative HIV-1 enzyme immunoassay is a surrogate marker for HIV-1 replication [13]. We assessed for changes in HIV-specific antibody titers as a component of the primary outcome measures. Significant decreases in antibody titers during the treatment period were observed only in participants in arm C (P = .0152, median decrease 32.19 [0.93, 63.45]) and arm D (P = .0332, median decrease 29.65 [11.56, 47.74]; Supplementary Figure 6B).

Following approval by the study steering committee, participants were offered the option to undergo ATI to evaluate posttherapy control and VL dynamics. As stated above, 2 participants (P24 in arm E and P26 in arm F) prematurely interrupted ART before the study steering committee approved that an ATI be pursued. In addition, individuals P17 and P19 (arm D) both opted not to undergo ATI. Ultimately, 26 of 30 participants successfully completed the ascribed interventions and consented to ATI. Following ATI, 25 participants rebounded within 14 weeks. No adverse events were observed during the ATI period.

Remarkably, participant P13, who was in the nicotinamide intensified regimen (arm C), maintained an undetectable plasma HIV VL with no blips up to 78 weeks post-ATI. This participant had the first detectable HIV VL (3.8 Log₁₀ copies/mL) at week 84 post-ATI and restarted ART at week 88 following resumption criteria. Two participants in arm F, P27 and P29, who received all combined therapies, maintained VLs that were intermittently below the level of detection or detectable but below 1000 copies/mL for 22 weeks post-ATI when ART was resumed as per participants’ decision (Figure 4A). Before ART was resumed, seminal plasma VL was measured in these 2 participants to evaluate the risk of HIV sexual transmission. Viral control was found to be undetectable (P27: <80 copies viral RNA/mL; P29: <40 copies viral RNA/mL) (differences in the detection limit are due to the volume of seminal plasma obtained), thus supporting the suppression of VL displayed by these 2 individuals and reduced potential risk of HIV sexual transmission [14, 15]. In both cases, ART was reinstated at week 22, when plasma VL was 299 and 514 copies/mL (P27 and P29, respectively) (data not shown). Notably, both participants’ CD4+ T-cell counts remained stable during the ATI period (Supplementary Figure 7).

Figure 4.

Viral loads during ATI. Individual plasma HIV-1 RNA levels are shown and coded by intervention arm (A). Viral load set point pre- and posttherapy values were based on the latest documented viral load measurement across the interventions (B). Viral load set point differences per group were evaluated by paired t-test (2-sided) following Log transformation.

We next assessed the relationships between peripheral or rectal tissue HIV DNA levels and viral rebound kinetics at pre- and postintervention time points. No significant correlations between time to VL rebound and viral DNA levels were observed across time points (per-protocol analysis; Supplementary Figure 8A–D).

Next, we assessed changes in VL set point pre- and posttherapy based on the latest documented VL measurements across the interventions. No significant changes in VL set points were observed across arms A through E (Figure 4B). However, in arm F, VL set point trended to decline posttherapy (per-protocol analysis; mean difference: −1.242 Log RNA copies/mL; 95% CL: −2.855 to 0.3710; P = .0873). As compared to the other interventional arms, group F displayed a trend toward a VL set point that was lower by 0.9534 Log ± 0.4806 Log HIV-1 RNA copies than those exposed to the incomplete interventions (per-protocol analysis; mean difference: −0.9534 ± 0.4806; 95% CL: −1945 to 0.03852; P = .0588), thus confirming the trend emerging from the comparison with the pretherapy values (Supplementary Figure 9A). Interestingly, when groups were stratified based on being on the different interventions, those on a nicotinamide regimen had a significant decline in post VL set point compared to those on a nonnicotinamide regimen (mean difference: −0.7469 Log RNA copies/mL; 95% CL: −1.472 to −0.02135; P = .0441) (Supplementary Figure 9B). However, after adjustment for multiple comparison, the significance was lost (P = .1323). No differences were seen in VL set point between those on the auranofin and nonauranofin regimens or between those on the DC versus the non-DC therapy (Supplementary Figure 9C and 9D). Furthermore, peripheral and rectal HIV DNA levels did not predict or associate with plasma VL set point (per-protocol analysis; Supplementary Figure 8E–H).

In search of possible host correlates of viral suppression, we next investigated widely utilized epigenetic biomarkers of immune aging. Accelerated epigenetic aging, based on epigenetic clock algorithms constructed from dynamic DNA methylation states, is observed in PWH, and early ART initiation during acute HIV may only, in part, reverse this process [16]. To determine whether immune cellular epigenetic age was impacted by multimodal treatments and immune-based interventions, we captured all participants’ longitudinal PBMC DNA methylation profiles before and after intervention [16]. We observed no significant differences between baseline and week 48 time points in epigenetic age estimates (PhenoAge Clock; Supplementary Figure 10A), mortality prediction (GrimAge Clock; Supplementary Figure 10B), and telomere length (DNAmTL; Supplementary Figure 10C) for all 6 study arms. Notably, only 1 participant (P13 in arm C), who also presented with the longest post-ATI control, exhibited a potential reversal of epigenetic age of approximately 15 years (estimated 42.8 years at baseline to 29.2 years at week 48). Additionally, this participant showed decreased epigenetic mortality risk and increased inferred telomere length. Interestingly, among the 2 posttreatment controllers in arm F, P27 showed a marked decline in GrimAge and P29 had a marked increase in inferred telomere length.

DISCUSSION

Our multi-interventional clinical trial revealed that combined candidate anti-HIV reservoir therapies across the 6 arms used in this study were deemed clinically safe in all participants and were associated with postintervention VL suppression of 3 participants who completed the study. Anecdotal cases of HIV cure have so far been mainly described following allogeneic bone marrow transplantation, which substitutes the individuals’ immune system encompassing the HIV reservoir cells [17]. In the multi-interventional arm F of our study, there was a trend in decreased total HIV DNA and markers of T cell activation in blood and, more importantly, a substantial reduction in post-ATI VL set point, an event so far rarely reported following scalable experimental anti-HIV reservoir interventions. The present study however has some limitations, including the use of total instead of integrated or intact proviral DNA as a surrogate marker of the viral reservoir and the low number of participants rendering it difficult to attribute causation to the single or combined agents adopted.

Nicotinamide, a NAD+/NADP+ precursor, is a sirtuin inhibitor known to inhibit HDAC class III through competitive binding to the sirtuin NAD+ binding site [18]. This trial was initially conceived to use nicotinamide as a latency reversal agent, given the published results on its effects on HIV latency [8]. In addition, it is known that HIV-1–infected cells display decreased NAD+ levels (the “intracellular pellagra” hypothesis) that can be corrected by nicotinamide supplementation [19]. The viral rebound dynamics we observed after ATI in participants in arms C and F may have suggested that nicotinamide contributed to this outcome; however, the significance was lost after multiple comparison adjustments; thus, the impact of nicotinamide will require further validation. Recent studies show that inhibition of lymphocyte apoptosis was shown to occur at drug concentrations achievable in vivo during nicotinamide supplementation [20 ]. Nicotinamide can also increase cytotoxic and helper T cell responses [21] and thus may have contributed to better immune-mediated control of residual HIV replication. Interestingly, nicotinamide decreases malignant microinflammation, linked to immune senescence and lymphocyte apoptosis [9]. Notably, 2 out of 3 individuals who displayed suppressed VLs post-ATI (all having received nicotinamide) also showed some level of reversal in epigenetic immune aging. This is intriguing given that nicotinamide and related molecules are being pursued as investigational senescence-reversing agents [22]. Such an effect will require further larger proof-of-concept studies.

Auranofin may contribute to the restriction of the viral reservoir through induction of central memory CD4+ T-cell differentiation to short-lived effector-like phenotype [23] as well as the induction of viral escape from latency [24]. Previous studies conducted in macaques had shown that auranofin, added to intensified ART regimens, contributed to a posttherapy decrease in VL set point [6]. Participants receiving auranofin (arm D) experienced a transient decrease in the viral reservoir compared to those on ART without auranofin [7]. However, there were no differences in viral rebound dynamics compared controls (arms A and B). The impact of auranofin may be limited.

Only the multiple combination intervention arm including DC therapy was associated with a decreased trend in viral set point post-ATI, with participant P27 demonstrating posttreatment controller features. Dendritic cell therapy is aimed at stimulating CD4+ and CD8+ T-cell responses. Our personalized approach was based on a calculated “perfect match” between the autologous virus-derived antigen and the individual's HLA. This might have helped to induce TCR stimulation of antigen-specific T-lymphocytes and their consequent clonal expansion. This intervention was the last in time that was administered to the study participants. This timing was aimed at ensuring that a sufficient number of clonally expanded antigen-specific T-lymphocytes might still be present during viral rebound following ATI. The expanded antigen-specific clones are expected to recognize and kill the productively HIV-1–infected cells following the reactivation from latency leading to viral rebound. Their activity in the multi-intervention group might have been enhanced by the previous use of both auranofin and nicotinamide. Auranofin, which is a prooxidant drug, was shown to increase immune effector responses through compensatory activation of the master antioxidant gene product, Nfr2. Nicotinamide is a precursor of NADH, a cofactor that is fundamental in the antioxidant response and the production of which is enhanced by Nrf2 Noteworthy, participants in arm F also displayed higher IFN-γ HIV Gag T-cell ELISPOT responses to the treatment peptides as compared to arm E participants. However, given the small sample size, we cannot rule out that factors other than the interventions (ie, genetic landscape) may have contributed to the study outcomes.

The combined treatment of arm F participants with drugs that may boost the effector response represents a difference between our approach as compared to a previous attempt at personalized immunotherapy for PLW (ACTG A5197). While a direct comparison of the outcomes is difficult, as the ACTG A5197 study enrolled a much larger number of participants than the present study, the results of both trials converge in highlighting that patient subgroups displaying the best IFN-gamma responses are also the ones showing an impact on the viral set point post-ATI.

The outcome of our proof-of-concept clinical trial resulted in 1 individual (P13), who received both ART intensification and nicotinamide, and 2 participants (P27, P29) in the multi-interventional group presenting with long-term viral suppression post-ATI. Participant P13 satisfied the definition of a posttreatment controller (ie, 2/3 of plasma VL measures <400 copies/mL through 6 months [25 ]). However, while P13 presented with a HIV protective HLA allele phenotype, his post ATI viral load set point was two orders of magnitude lower than his pre-ART viral load set point suggesting a possible role of the interventions in this control. Participants P27 and P29 who both lacked any HIV protective HLA alleles, approximated the posttreatment controller definition having remained off ART for 22 weeks and matched the status of viral load suppression according to the World Health Organization's new guidelines on viral control (VL <1000 copies/mL) [26]. Participant P12, also from arm C (ART intensification plus nicotinamide), presented a VL of 550 copies/mL at week 39 days post-ATI, however decided to resume ART without further evaluation for possible virologic control. Furthermore, participants P27 and P29 matched the definition of viremic controllers [27], presenting with plasma VLs post-ATI consistently under 5000 copies/mL through this study period. These post-ATI dynamics support the redesign of future larger clinical trials to dissect the contribution of each intervention toward posttherapy HIV control.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.