Altered Immune Reconstitution in Allogeneic Stem Cell Transplant Recipients With Human Immunodeficiency Virus (HIV)

Abstract

Background

Persons living with human immunodeficiency virus (HIV) are at elevated risk of developing the malignant diseases that require allogeneic stem cell transplantation (ASCT). Recent data suggest that these individuals are also at an elevated risk of certain complications post-ASCT. This risk may result from preexisting HIV-related factors affecting dynamics of immune reconstitution post-ASCT. However, to date, there has been little work describing the dynamics of immune reconstitution post-ASCT in persons with HIV and none comparing these data to controls without HIV.

Methods

We assessed T-cell reconstitution in 6 ASCT with HIV recipients (HIV+ASCT) compared to a control population of 21 ASCT without HIV recipients. In a subset of HIV+ASCT recipients we performed additional flow cytometry profiling of CD8+ T-cell subsets and antigen specificity of reconstituting CD4+ and CD8+ T cells.

Results

We observe no difference in post-ASCT CD4+ T cells between HIV+ASCT and HIV-negative ASCT recipients, despite much lower pre-ASCT CD4+ T-cell counts in the HIV+ASCT group. In contrast, we observed significantly higher CD8+ T-cell numbers in the HIV+ASCT group post-ASCT. The reconstituting CD8+ T-cells were predominantly CD45RO+, whereas homing markers and antigen specificity of these cells varied between participants.

Conclusion

This study represents the most extensive characterization of immune-reconstitution post-ASCT in persons with HIV, and the first to our knowledge to compare these data to ASCT controls without HIV. The results indicate that immune reconstitution in this group can be affected by preexisting HIV infection and post-ASCT antigen exposure.

We describe T-cell reconstitution postallogeneic stem cell transplantation in 6 people with human immunodeficiency virus (HIV) compared to 21 controls without HIV. We observed a significantly elevated CD8+ T-cell expansion post-Allogeneic Stem Cell Transplant in recipients with HIV. The nature and specificity of the CD8+ T-cell population differed among individuals.

Interest in the role of allogeneic stem cell transplants (ASCT) in persons living with human immunodeficiency virus (HIV) on antiretroviral therapy (ART) has been increasing for clinical and scientific reasons. First, persons with HIV are at increased risk of developing certain malignancies for which ASCT is an established treatment modality [1–3]. In addition, following the report of HIV cure in 2 individuals post-ASCT [4, 5], there is substantial interest in the possible effects of ASCT on the latent HIV reservoir.

Although the safety of ASCT in carefully selected persons with HIV is generally accepted based on case series, some reports suggest an increased risk of certain infections, including cytomegalovirus (CMV) infection and nontuberculous mycobacterial infection [6]. One possible explanation of such an increase in infection risk would be a perturbation of posttransplant immune reconstitution in persons with HIV, potentially resulting from the preexisting immune deficiency and associated factors, including fibrosis of lymphoid architecture [7, 8].

In individuals without HIV, the dynamics of immune reconstitution post-ASCT varies. Granulocyte recovery is clearly characterized and typically occurs early (within 21 days post-ASCT). Mononuclear cell recovery is later, with numerical recovery of natural killer cells and T cells generally complete around 100 days post-transplant [9], although the timing of functional recovery is less clear. Reconstitution of CD4+ T cells is dependent on thymic production of naive T cells and generally occurs after the recovery of CD8+ T cells. B cells generally recover much later, up to 2 years post-ASCT [9]. The timing and tempo of T-cell reconstitution is particularly important for mediating important clinical events, including infections, graft-versus-host disease (GVHD), and relapse.

Characterization of immune reconstitution in persons with HIV post-ASCT is limited, with no data comparing this group to individuals without HIV post-ASCT. Therefore, in this study, we assessed CD4+ and CD8+ T-cell reconstitution post-ASCT in a cohort of 6 persons with HIV all on suppressive ART, in comparison with a control cohort of ASCT recipients without HIV. We further characterized cellular subsets, including trafficking and differentiation markers, and the kinetics of reconstituting T-cell responses to pathogen-specific antigens in the group with HIV.

METHODS

Six ASCT recipients with HIV (HIV+ASCT) (Pt 1–6) were recruited from parallel studies at the Kirby Institute/St Vincent’s Hospital Sydney, the Alfred Hospital Melbourne, The Royal Melbourne Hospital and the Doherty Institute, University of Melbourne. ASCT recipients without HIV (ASCT controls) were recruited from a contemporaneous parallel study at St Vincent’s Hospital, Sydney. An additional healthy control (HC) cohort, consisting of 10 non-ASCT recipients without HIV, was recruited as part of an internal laboratory quality control study. All individuals gave written informed consent for collection and analysis of samples. All studies were approved by the relevant institutional Human Research Ethics Committees/Institutional Review Boards. The HIV-reservoir characteristics of Pt-3, Pt-4, and Pt-6 have been reported previously [10].

Prospectively collected T-cell counts were analyzed at the closest clinically practical time points to 12 months post-ASCT in both groups. Additional flow cytometric characterization was then performed on cryopreserved peripheral blood mononuclear cells (PBMCs) from 3 HIV+ASCT participants; cryopreserved PBMCs were not available in the other participants. Detailed analysis of CD8+ lymphocyte populations was performed as previously described [11] and analyzed in comparison with baseline data from the healthy non-ASCT control cohort performed using the same methods.

Intracellular cytokine staining (ICS), as described previously [12], was used to assess antigen specificity of reconstituted subsets to the mitogen staphylococcal enterotoxin B (SEB), and to pools of peptides from key recall antigens, including CMV pp65 and IE-1, EBV (lytic and latent proteins), and human immunodeficiency virus type 1 Gag and Nef (Miltenyi Biotec, Germany, and National Institutes of Health AIDS Reagent Program). Flow cytometric data were analyzed using FlowJo (v10.4.2; TreeStar, Ashland, OR, USA). For ICS, background responses from the no antigen tube were subtracted from all other antigen tubes. A positive response for any cytokine was defined as greater than twice the background and greater than 20 T-cell events.

Statistical comparisons were performed in GraphPad Prism (GraphPad, La Jolla, CA, USA). Comparisons of total lymphocyte counts were performed using a Mann-Whitney U test with results considered significant with P < .05.

RESULTS

Clinical Characteristics and Clinical Course post-ASCT

Baseline clinical characteristics and the clinical course post-ASCT are detailed in Table 1. Additional patient specific characteristics for the HIV+ASCT group can be found in Supplementary Tables 1–3. Median ages of the HIV+ASCT and ASCT-control groups were similar (50 vs 48), whereas fewer HIV+ASCT individuals underwent myeloablative conditioning compared to ASCT controls (1/6 vs 14/21). The HIV+ASCT group had a higher proportion of males (5/6 vs 13/21). All HIV+ASCT recipients were on ART with full viral suppression. Median pre-ASTC CD4+ T-cell counts were slightly lower for the HIV+ASCT group (267 vs 368).

Table 1.

Clinical Characteristics for All Participants

	HIV+ASCT (n = 6)	ASCT Controls (n = 21)
Demographics:
Age (median, range)	50, 42–60	48, 23–68
Male	5/6	13/21
Race (no. white)	5/5a	n/a
Underlying malignancy:
Acute myeloid leukemia/myelodysplastic syndrome	4/6	13/21
Plasmablastic non-Hodgkins	2/6	2/21
Chronic lymphocytic leukemia	0/6	2/21
Aplastic anemia	0/6	1/21
Acute leukemia (otherwise unspecified)	0/6	2/21
Acute lymphoid leukemia	0/6	1/21
Conditioning:
Myeloablative	1/6	15/21
Reduced intensity	5/6	6/21
HIV-specific characteristics:
Pre-ASCT CD4+ T-cell count (median, range) cells/µL	267, 48–789	368, 170–1603
On ART pre-ASCT	6/6	n/a
Pre-ASCT HIV viral load <50 (pr < 20) copies/mL	6/6	n/a
Transplant-related outcomes:
1 year TRM	0/6	0/21
100d TRM	0/6	0/21
CMV reactivation	2/6	1/21
Participants diagnosed	Pt-3 and Pt-5	n/a
Acute GVHD post-ASCT	3/6	9/21
Participants diagnosed	Pt-1, Pt-3, and Pt-4	n/a
Chronic GVHD post-ASCT	3/6	12/21
Participants diagnosed	Pt-3, Pt-5, and Pt-6	n/a
Relapse	0/6	3/21

Abbreviations: ART, antiretroviral therapy; ASCT, allogeneic stem cell transplantation; CMV, cytomegalovirus; GVHD, graft-versus-host disease; HIV, human immunodeficiency virus; n/a, not available; Pt, patient; TRM, transplant-related mortality.

^aEthnicity data not available for 1 participant.

Post-transplant clinical progress was broadly similar between HIV+ASCT and ASCT-control groups. All participants achieved ≥99% donor chimerism post-ASCT. Clinical progress, assessed by transplant-related mortality (TRM) (zero in both groups), CMV reactivation (2/6 and 1/21), acute GVHD (3/6 and 9/21), and chronic GVHD (3/6 and 12/21) was generally comparable, although the comparison is limited by the small HIV+ASCT population and the divergent conditioning. Among the HIV+ASCT group, Pt-3 and Pt-5 were diagnosed with CMV reactivation post-ASCT. Pt-1, Pt-3, and Pt-5 were diagnosed with acute GVHD post-ASCT. Pt-3, Pt-5, and Pt-6 were diagnosed with chronic GVHD post-ASCT.

In the HIV+ASCT group, all persons remained supressed post-ASCT except for Pt-3. This participant experienced several instances of HIV-rebound post-ASCT (Supplementary Figure 1), despite reporting full ART adherence. At time of the first viral rebound the patient had reduced drug levels [10], likely due to poor drug absorption caused by gastrointestinal GVHD as this persisted following extensive efforts to maximize adherence.

Immune Reconstitution in Peripheral Blood

We first compared CD4+ and CD8+ T-cell counts between HIV+ASCT and ASCT-controls at time-points close to 12 months post-ASCT. There were no significant differences in CD4+ T-cell counts between HIV+ASCT (median = 367 cells/µL [interquartile range (IQR) = 283–471]) and ASCT controls (313 cells/µL [176–620]) (P = .74) (Figure 1A). In contrast, CD8+ T cells were significantly higher in HIV+ASCT (1676 cells/µL [670–2226]) compared to ASCT-controls (550 cells/µL [328–1280]) (P = .049) (Figure 1A), with the CD8+ T cells continuing to rise after 12 months and persisting for several years post-transplant in HIV+ASCT recipients (Figure 1B). All available CD4+ and CD8+ T-cell counts from time-points pre and post-ASCT for the HIV+ASCT group is reported in Supplementary Table 2. The counts for the ASCT-controls are reported in Supplementary Tables 3 and 4.

Figure 1.

A, CD4+ T cells and CD8+ T cells, measured between 7 and 12 months post-ASCT, of HIV+ASCT and ASCT controls. There were no differences between the CD4+ T-cell counts of HIV+ASCT (median = 367 cells/µL [IQR = 283–471]) and ASCT controls (313 cells/µL [176–620]). However, CD8+ T cells were significantly higher in HIV+ASCT subjects (1676 cells/µL [670–2226]) compared to ASCT controls (550 cells/µL [328–1280]). B, Longitudinal CD8+ T-cell counts for available measurements in HIV+ASCT participants. Abbreviations: ASCT, allogeneic stem cell transplantation; HIV, human immunodeficiency virus; IQR, interquartile range.

Detailed Characterization of CD8+ T-Cell Subsets

We next characterize CD8+ T-cell subsets of the 3 participants (Pt-1, Pt-2, and Pt-5) who had cryopreserved PBMCs available pre- and post-ASCT. Pt-1 had an expansion of CD45RO+ CD8+ T cells post-ASCT, consisting mainly of integrin-α4(CD49d) + integrin-ß7- non-gut/mucosal-homing cells (Figure 2 and Supplementary Table 5). Pt-2 also had an expansion of CD45RO+ CD8+ T cells post-ASCT, consisting mainly of integrin-α4(CD49d) + integrin-ß7+ gut/mucosal homing effector memory cells (Figure 2 and Supplementary Table 5). There was no clear elevation of CD8+ T-cell counts in Pt-5.

Figure 2.

A, Flow cytometric characterization of CD8+ T-cell subsets post-ASCT in 3 HIV+ASCT individuals. Pt-1 predominantly had an elevation of CD45RO+ non-gut homing CD8+ T-cells post-ASCT. Pt-2 had an elevation of CD45RO+ activated (CD38+) gut-homing CD8+ T cells, particularly at 7–8 months post-transplant. No clear expansion of CD45RO+ CD8+ T cells was observed in Pt-5; therefore, CD38+ data are not presented for this patient. Subsets are presented as percentages of total CD8+ T cells. B, Flow cytometric data on CD8+ T cells compared to a group of HCs. The ASCT with HIV group exhibited a much higher percentage of memory CD45RO+ CD8+ T cells and highly activated CD38+CD127- CD8+ T cells than HCs. The data from the ASCT with HIV group came from time-points 7, 8, and 9 months post-ASCT for Pt-1, 2, and 5, respectively. Abbreviations: ASCT, allogeneic stem cell transplantation; HC, healthy control; HIV, human immunodeficiency virus; Pt, patient.

When compared to the healthy non-ASCT controls, (Figure 2B and Supplementary Table 6), CD8+ T cells (as a percentage of CD3+ T cells) were elevated in ASCT recipients with HIV (median of 33% in the healthy control (HC) vs 79% in ASCT with HIV). Of these cells, memory CD8+ T cells were overrepresented in the ASCT group with HIV (26% vs 47%). These memory CD8 cells were also highly elevated as a percent of CD3+ T cells (8% vs 38%). Also in contrast with healthy non-ASCT controls, a majority of these memory CD8 T cells were highly activated, CD127- CD38+ (10% vs 75% as a percentage of CD8+ T cells and 0.9% vs 25% as a percentage of CD3+ T cells; Supplementary Figure 2).

Antigen Specificity of Reconstituting T Cells

Next, we assessed the antigen specificity of CD4+ and CD8+ T cells in the same 3 participants (Pt-1, Pt-2, and Pt-5) pre- and post-ASCT. Background subtracted raw values for responses to all antigens are listed in Supplementary Tables 7 and 8.

Pt-1 had a small (<10% of SEB response) response to CMV antigens post-ASCT. There was also a small CD8 + IFNγ+ and CD4 + IFNγIL-2+ response to HIV-Nef, 4 months post-ASCT.

Pt-2 had small, inconsistent CD8+ responses to EBV post-ASCT. However, there was a very clear response to CMV antigens at all time-points post-ASCT (Supplementary Tables 7 and 8). This response was characterized by very large (>50% of SEB response) CD8+ and CD4+ responses to CMV pp65 (3 months post-ASCT) and CMV IE-1 (8, 14, and 21 months post-ASCT). Pt-2 also had small CD8 + IFNγ, CD8 + TNFα+ and CD4 + IFNγ + TNFα+ responses to HIV-Nef 8 and 14 months post-ASCT.

Pt-5 also exhibited a variety of large (10–50% of SEB response) CD4+ and CD8+ T-cell responses to all tested CMV antigens at all time-points post-ASCT, particularly to CMV pp65 9 months post-ASCT. There were also CD4+ and CD8+ responses to EBV at all time-points post-ASCT. Pt-5 did not have any observable responses to HIV antigens.

DISCUSSION

In this study we assessed CD4+ and CD8+ T-cell reconstitution post-ASCT in a group of HIV+ASCT recipients, compared to controls without HIV. We observed similar levels of CD4+ T-cell reconstitution post-ASCT in the 2 groups but significantly higher CD8+ T-cell reconstitution in the HIV+ASCT group. These higher CD8+ T-cell numbers occurred between 9 and 12 months post-ASCT and persisted for several years post-ASCT. Furthermore, among a subset of HIV+ASCT recipients who had post-ASCT cryopreserved PBMCs available for analysis, we saw that the expanded CD8+ T cells post-ASCT were heterogenous with regards to both the homing markers present on the cell surface and the antigen specificity of the cells.

The main strength of this study was the ability to compare ASCT recipients with HIV and without HIV. By comparing these groups, we first observed that there was no clear difference in the post-ASCT CD4+ T cells between the HIV+ASCT and the groups without HIV. This observation is somewhat surprising given that the HIV+ASCT group had a lower CD4+ T-cell count pre-ASCT and the propensity for HIV infection to damage lymphoid structures key to the proliferation of CD4+ T cells [7, 8]. Importantly, these results suggest that HIV+ASCT recipients are able to mount a recovery of CD4+ T cells comparable to their counterparts without HIV. In contrast, we saw that the CD8+ T cells are higher in the HIV+ASCT group than ASCT controls. Although data regarding CD8+ T-cell recovery is generally limited in studies in this population, comparable evidence of elevated CD8+ T cells post-ASCT was recently reported in the “London patient” [5], who exhibited HIV clearance post-ASCT. It should be noted that one of the participants in our study (Pt-3) experienced several incidences of HIV rebound despite having elevated CD8+ T cells post-ASCT, suggesting that the finding of CD8+ T-cell elevation is not broadly associated with HIV clearance post-ASCT.

To further elucidate the subsets and antigen specificity of the elevated CD8+ T cells in the HIV+ASCT recipients, we performed further analyses in a subset of patients who had available cryopreserved PBMCs These analyses revealed expansions of CD45RO+ CD8+ T cells in Pt-1 and Pt-2 (both with elevated CD8+ T-cell counts post-ASCT) but not in Pt-5, as similarly reported in a recent case study [13]. More detailed analysis of the CD45RO+ CD8+ T-cell subsets showed that they differed between HIV+ASCT recipients, with both integrin-α4(CD49d) + integrin-ß7- non-gut/mucosal-homing (Pt-1) and integrin-α4(CD49d) + integrin-ß7+ gut/mucosal homing (Pt-2) effector memory expansions observed. In the 2 participants that had pre-ASCT samples available, this skewing toward either gut or non-gut homing cells existed prior to transplant. This suggests that immune reconstitution may have been influenced by preexisting factors in these participants.

Alternatively, this variation in homing markers may reflect a dysregulated response to different antigens post-transplant in HIV+ASCT recipients. Characterization of CD8+ T-cell antigen specificity support this, with a very large CMV response observed, despite no observed clinical CMV infection, at the same time-points as the CD45RO + CD8+ expansion in Pt-2 but not Pt-1. This suggests a possible role for subclinical recurrent CMV in Pt-2 and perhaps another unidentified antigenic exposure in Pt-1. Additionally, EBV and HIV reactivity of reconstituting T cells were also heterogenous.

Interestingly, Nef-specific CD4+ and CD8+ T-cell responses post-ASCT were observed in both Pt-1 and Pt-2. Pt-1 had a small CD8+ (IFNγ+ and TNF+) response to Nef pre-ASCT and a small CD8+ (IFNγ+) and CD4+ (IFNγ + IL2+) response to Nef 4 months post-ASCT. Pt-2 also had a small CD8+ (IFNγ+ and TNF+) and CD4+ (IFNγ+TNF+) Nef-specific response 8 and 14 months post-ASCT. Pt-5 did not have any observed HIV-specific responses. This is reminiscent of the primary immune response to HIV infection (where Nef-specific CD4+ and CD8+ T cells are the first to develop [14]) and could indicate that the transplanted immune system is encountering HIV for the first time. Post-ASCT HIV-specific immune responses have been observed in other studies [13, 15, 16] and may indicate low level viral replication or viral protein production, despite undetectable HIV viral load.

Our analyses were limited by the small cohort size, inevitable given this rare and difficult to study population, some variation in time-points and conditioning across patient groups, and the retrospective nature of the study. We were also limited by the lack of available PBMCs in the ASCT without HIV cohort with which to perform detailed analyses of immune cell subsets and antigen specificity. Future studies should aim to further characterize immune-reconstitution and associated clinical outcomes in this rare but increasingly important group of HIV+ASCT patients, compared to well-matched controls using a standardized prospective study design.

In conclusion, this study represents the most detailed description of immune reconstitution post-ASCT in persons with HIV and the first comparison to controls without HIV. However, the differences seen in this study nonetheless suggest that immune reconstitution and activation post-ASCT can be affected by preexisting HIV infection. Mechanisms of this alteration could include certain coinfections, prior disruption of immune signaling pathways, preexisting damage to lymph-node architecture, or ongoing production of immune-stimulatory HIV proteins. Although the HIV+ASCT recipients did not exhibit excessive clinical complications post-ASCT, further evaluation of any effect this CD8+ T-cell expansion is having on transplant-related clinical outcomes (eg, GVHD and infectious complications) is warranted in additional cohorts. Overall, these data highlight how little we know about the clinical course post-ASCT for persons with HIV and suggest that immune reconstitution in these individuals is sensitive to antigen exposure post-ASCT.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Author contributions. D. D. M., S. R. L., A. D. K., and M. N. P. designed the research. J. Z., C. M. L. M., C. F., and M. V. performed the experiments and analyzed experimental results. S. T. M., S. A., J. S., J. K., S. D., K. K., and S. R. L. contributed to the clinical cohorts described in the paper. D. D. M. and M. L. performed statistical analyses. D. D. M. made the figures. D. D. M., J. Z., and M. N. P. wrote the paper. All authors contributed to the editing of the paper.

Acknowledgments. The authors thank the late David A. Cooper for his guidance and support throughout the course of this study. Vale David. The authors also thank the study participants for donating their time and samples for this study. We also acknowledge the contributions of Annabelle Horne and Jason Gao at St Vincent’s Hospital Sydney, as well as Drs Nenad Mecasic, Christina Chang, and Thomas Rasmussen for assistance in the protocol development and ethics application at The Alfred and Royal Melbourne Hospitals and Julia Stout, Jennifer Audsley, and Barbara Scher for administrative and ethics support of the HIV Cure Program at the Doherty Institute.

Financial support. This work was supported by the Australian National Health and Medical Research Council (NHMRC) (program grant 1052979) to A. D. K., S. R. L.). J. M. is funded by the St Vincent’s Clinical Foundation. S. R. L. is funded by the Delaney AIDS Research Enterprise (DARE), National Institutes of Health (U19 AI096109 and UM1AI126611) and a high impact research grant from the American Foundation for AIDS Research (109226-58-RGRL). S. R. L. and M. N. P. receive NHMRC Fellowships (1042654). M. N. P. is also supported by the Cancer Institute NSW Future Research Leader Fellowship. D. D. M. is supported by a postdoctoral fellowship from the Danish National Research Fund (DNRF126).

Potential conflicts of interest. M. P. holds a patent for immumodulatory agents in viral malignancies. S. R. L. reports grants to her institution from Gilead Sciences, Merck, Viiv, and Leidos, outside the submitted work. M. L. reports grants from Gilead Sciences, Janssen-Cilag, and Viiv Healthcare, outside the submitted work. All other authors declare no conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.