Mesenchymal stromal cells in human immunodeficiency virus‐infected patients with discordant immune response: Early results of a phase I/II clinical trial

Abstract Between 15% and 30% of HIV‐infected subjects fail to increase their CD4+ T‐cell counts despite continuous viral suppression (immunological nonresponders [INRs]). These subjects have a higher morbidity and mortality rate, but there are no effective treatments to reverse this situation so far. This study used data from an interrupted phase I/II clinical trial to evaluate safety and immune recovery after INRs were given four infusions, at baseline and at weeks 4, 8, and 20, with human allogeneic mesenchymal stromal cells from adipose tissue (Ad‐MSCs). Based on the study design, the first 5 out of 15 INRs recruited received unblinded Ad‐MSC infusions. They had a median CD4+ nadir count of 16/μL (range, 2‐180) and CD4+ count of 253 cells per microliter (171‐412) at baseline after 109 (54‐237) months on antiretroviral treatment and 69 (52‐91) months of continuous undetectable plasma HIV‐RNA. After a year of follow‐up, an independent committee recommended the suspension of the study because no increase of CD4+ T‐cell counts or CD4+/CD8+ ratios was observed. There were also no significant changes in the phenotype of different immunological lymphocyte subsets, percentages of natural killer cells, regulatory T cells, and dendritic cells, the inflammatory parameters analyzed, and cellular associated HIV‐DNA in peripheral blood mononuclear cells. Furthermore, three subjects suffered venous thrombosis events directly related to the Ad‐MSC infusions in the arms where the infusions were performed. Although the current study is based on a small sample of participants, the findings suggest that allogeneic Ad‐MSC infusions are not effective to improve immune recovery in INR patients or to reduce immune activation or inflammation. ClinicalTrials.gov identifier: NCT0229004. EudraCT number: 2014‐000307‐26.


| INTRODUCTION
Chronic immune activation and inflammation is considered today as the main driving force of CD4 + T-cell depletion and the functional impairment of the immune system caused by HIV infection. 1 Antiretroviral therapy (ART) achieves the control of viremia in most subjects and reduces both cellular and soluble activation markers, leading to immune recovery in a high proportion of subjects. 2 However, between 15% and 30% of subjects exhibit a poor CD4 + T-cell recovery despite successful viral suppression; these subjects are known as immunological nonresponders (INRs). 3 Although the definition of immunological nonresponse lacks consensus, it has always been based in the increase of CD4 + T-cell counts above different thresholds in a given time period. 4 These subjects show severe homeostatic alterations in CD4 + T cells, with a disturbed maturational profile, a reduced thymic function, and increased levels of activation and apoptosis, among other characteristics. 5 From a clinical point of view, INRs show higher rates of morbidity and mortality associated with AIDS and non-AIDS events such as cardiovascular events, neurocognitive impairment, non-AIDS malignancies, end-stage liver and renal diseases, bone disorders, and frailty than those with a good immune response. [6][7][8] Moreover, when these subjects grow old, this situation will be aggravated by age-associated immunosenescence. 9 In these subjects, many strategies have been evaluated, such as ART intensification, immunomodulators, immunosuppressive agents, and probiotics, though with disappointing results; thus, no current effective therapies are available. 10 On the other hand, several studies, both in vitro and in vivo, have shown that mesenchymal stromal cells (MSCs) can modulate the function of T helper cells and B lymphocytes, natural killer (NK) cells, and dendritic cells, whereas stimulating regulatory T (T reg ) cells results in a change from a proinflammatory state to an anti-inflammatory state. [11][12][13] These properties have been demonstrated in multiple animal models of disease and have been used successfully in humans with graft vs host disease and several autoimmune and nonimmune diseases. 14,15 Up to now, only one study has been carried out with MSCs from cord blood in INRs, which resulted in a significant increase in circulating CD4 + T lymphocytes and a decrease of the activation of T lymphocytes and soluble inflammation mediator levels without significant adverse effects or loss of viremia control, 16 but this study has not been replicated. Thus, our aim was to evaluate, for the first time, whether MSCs coming from a more accessible source, such as adipose tissue, are safe and effective in improving the immune recovery in INRs.

| Study design
This was originally planned as a phase I/II, randomized, placebo-controlled, clinical trial designed to evaluate the safety and efficacy of adipose tissue allogeneic adult MCSs (Ad-MSCs) in INRs. The design was carried out jointly with the Andalusian Network for the Design and Translation of Advanced Therapies (http://terapiasavanzadas.juntaandalucia.es), which also acted as sponsor.
Because of security concerns, in the first phase, five eligible INRs received unblinded Ad-MSCs with a safety minimum period of 15 days between patients. Once all five subjects had completed the four Ad-MSC infusions, an independent data-monitoring committee (IDMC) performed a preliminary analysis of safety and efficacy data. In the second phase, (permanently suspended), 10 additional subjects would have been randomized to receive Ad-MSC infusions or placebo.

| Study approval
The clinical trial was approved by the National Health Authority and the Ethics Committee for Clinical Research of the participating site.

Lessons learned
• Adipose tissue allogeneic adult mesenchymal stromal cells infusions are not effective to improve immune recovery or to reduce immune overactivation or inflammation state in immunologically nonresponding HIVinfected patients.
• Donor-associated factors and manufacturing procedure could affect efficacy and safety.  (Table S1). Once the adipose tissue was obtained, a mechanical disintegration of the tissue was performed, followed by an enzymatic digestion with collagenase type A. The cell fraction was separated by centrifugation and seeded in plates, and after two culture-expansion passages Ad-MSCs were isolated. The formulation of medium for Ad-MSC expansion was as follows: Dulbecco's modified Eagle's medium with 10% of fetal bovine serum, 2% of L-alanine and L-glutamine, 0.1 mg/mL of gentamicin, and 100 UI/mL of penicillin. After the expansion Ad-MSCs were frozen and put into quarantine until quality controls were performed (Table S2). When a patient was included in the clinical trial, Ad-MSCs were thawed and expanded for 1 week, and new quality controls were performed before delivery (Table S3)

| Cell infusion procedure
Ad-MSCs were administered through a peripheral venous catheter over 1 to 2 hours using an infusion pump (infusion rate 2 mL/min) at a dose of 1 × 10 6 Ad-MSCs per kilogram of body weight at baseline and at weeks 4, 8, and 20 ( Figure 1). Before administration, the cell suspension was tempered and stirred manually or using an electric stirrer to dissolve any cell aggregates that could have occurred during transport, and subjects received premedication with methyl prednisolone (0.5 mg/kg i.v.), dexchlorpheniramine (5 mg, i.v.), and oral acetaminophen.

| Study subjects
The subjects were selected among the HIV-infected adults followed at the Virgen del Rocío University Hospital. An INR was defined as an HIV-infected subject with a basal CD4 + T-cell count ≤350 cells per microliter whose CD4 + T-cell count increased <75 or <150 cells per microliter after 1 or 2 years with undetectable viral load, respectively, and/or a CD4 + T-cell count increase <350 cells per microliter after 3 years on treatment. Exclusion criteria included opportunistic infections in the previous 12 months, active coinfections with hepatitis B virus or hepatitis C virus, Child-Pugh class C liver cirrhosis, portal hypertension or hypersplenism, malignant tumors, or treatment in the previous 12 months with immunomodulators, interferon, chemotherapy, or any other drug that might alter CD4 + T-cell count. Pregnant or breastfeeding women and subjects refusing to use accepted contraceptive methods throughout follow-up were also excluded from participation in the trial. All subjects agreeing to participate in the clinical trial provided written informed consent before undergoing any studyrelated procedure.

| Endpoints, follow-up, and assessments
The main aims of the study were to assess the safety and efficacy of      (Table 1).

| Safety and tolerability of Ad-MSC infusions in HIV-infected subjects
In total, 10 venous thrombosis events, 24 to 48 hours after Ad-MSC infusion, were observed in three out of the five subjects in the arms F I G U R E 1 Scheme of visits, infusions of Ad-MSCs, and sampling time points. *, safety visits, 1 week before the next Ad-MSC infusion. Ad-MSC, adipose tissue allogeneic adult mesenchymal stromal cells where Ad-MSCs were infused, which required low molecular weight heparin treatment. A decrease in the emergence of thrombotic events was observed when low molecular weight heparin (80 mg of enoxaparin per day) was given 1 day before, on the day of the infusion, and during the following 2 days, together with a retraining course on the cell infusion procedure, as it was detected that the infusion rate was significantly higher than that established by protocol in one case. All other adverse events (n = 5) were considered unrelated to the Ad-MSCs (Table 2). Furthermore, no subjects presented alterations in the biochemical and hematological parameters or increases in HIV viral load (data not shown).

| Efficacy of Ad-MSC infusions in HIV-infected subjects
Overall, no significant changes were observed in the CD4 + T-cell counts, percentage of CD4 + , or CD4 + /CD8 + ratios after infusions and throughout follow-up ( Figure 2). Likewise, there were no significant changes in the different subsets of CD4 + and CD8 + T lymphocytes (T N , RTE, T CM , T EM , and T EMRA ) except for an increase in the T EM CD8 + T subset ( Figures S3 and   S4); nor were there changes in the percentage of NK cells, T reg cells, mDCs, and pDCs ( Figure S5). Furthermore, we did not observe changes in the activation, proliferation, senescence and apoptosis, or exhaustion markers of CD4 + or CD8 + T cells. We observed a decrease in the percentage of PD1 + CD4 + T cells (5.2 vs 1.6; P = .043) and a trend in PD1 + CD8 + T cells (0.9 vs 0.4; P = .080) at week 96 ( Figures S6 and S7). On the other hand, the sCD14 plasma levels, measured with monocyte activation markers and the different proinflammatory proteins (IL-6, TNF-α, hsCRP) showed no significant changes ( Figure S8). Likewise, the viral reservoir, measured as total cell-associated HIV-DNA, remained stable throughout the follow-up ( Figure S9).

| DISCUSSION
Among the different properties of MSCs, their immunoregulatory potential is noteworthy because they can interact with cells of both the innate and adaptive immune systems, leading to the modulation of several effector functions. 18,19 MSCs are able to suppress T lymphocyte activation and proliferation by decreasing the production of TNF-α and IFN-γ while inducing IL-10 and IL-4 expression by CD4 + T cells. In addition, they inhibit the dendritic cell maturation and natural killer cell activation, induce T reg cell differentiation, and promote a shift of macrophage toward antiinflammatory phenotype, as well as secrete anti-inflammatory cytokines such as IL-1Ra, IL-10, TGF-β, and hepatocyte growth factor, among others. 14,18,20 Moreover, given their low immunogenicity, both autologous and allogeneic cells can safely be administered. [21][22][23][24] Zhang et al. 16  In our study, the subjects received four doses (1 × 10 6 /kg) of Ad-MSCs, but no changes were observed in the CD4 + T-cell counts, percentages, or CD4 + /CD8 + ratios, and no consistent changes were observed in the different subsets and phenotypes of the CD4 + and CD8 + T cells or percentage of mDCs, pDCs, NK cells, or T reg cells.
Likewise, the infusions of Ad-MSCs had no effects on sCD14, IL-6, TNF-α, and hsCRP plasma levels. Overall, we did not find a decrease in activation, exhaustion, apoptosis, and senescence at week 96. Only a significant decrease in the PD1 + CD4 + T cells was found, but its meaning is uncertain, and this change did not influence CD4 + T-cell recovery.
The principal differences with our study are the origin of the MSCs and the doses administered. Whereas Zhang et al. 16  F I G U R E 2 Evolution of the CD4 + T-cell counts, percentages, and CD4 + /CD8 + ratios after 96 weeks of follow-up. IQR, interquartile range that genes associated with cell adhesion, proliferation, and modulation of the immune system are enriched in Wharton's jelly-derived MSCs. 28 Likewise, recently, it has been proposed that other factors, such as tobacco, diabetes, or morbid obesity, can influence not only cell performance during manufacturing but also their pharmacological action. 29,30 All this could have influenced the lack of efficacy of MSCs in our study.
The high number of venous thrombosis events (VTEs) observed could have a complex multifactorial causality. On one side, the study subjects may be at increased risk of VTEs because of the patients' history of parenteral drug use (2/5), higher risk of recurrence in cases of previous thrombotic episodes, 31,32 or a high infusion rate (>3 mL/min) recorded in some infusions. In fact, the total of 10 events occurred in only three subjects, all having received four doses of the Ad-MSC suspension. Furthermore, in one patient, the VTEs could have been caused by the use of small-caliber veins on the back of the hands for the infusions, as this patient could not be infused in the arm. In addition, it has been reported that MSCs express tissue factor and have procoagulant activity, being observed more frequently in Ad-MSCs than in those of other sources. Furthermore, cell dose, handling conditions, growth media, and donor-associated factors might also influence procoagulant activity. [33][34][35][36] Our study has several limitations. Regarding the safety of Ad-MSCs, it would have been desirable to determine their tissue factor expression before their administration, although in more than 150 patients treated with this same product in different clinical trials sponsored by the Andalusian Network for the Design and Translation of Advanced Therapies, thrombotic events occurred exceptionally.
Furthermore, a full characterization of the infused MSCs was carried out, but neither in vitro functional assay to determine their immunomodulatory potency nor biodistribution analysis was performed. However, previous studies have shown that after intravenous infusion, cells were accumulated in lung, spleen, liver, and bone marrow. 37 On the other hand, a higher number of patients and a control group, planned in the second phase of the trial, would have allowed us to better assess the fluctuations that some patients presented in several immunological parameters. Furthermore, the analyses are based on PBMC samples, which may not always reliably reflect tissuerelated processes during HIV infection. The use of lymph node samples instead of PBMCs would have needed periodical biopsies, lowering the feasibility of the study.

| CONCLUSION
The Ad-MSC infusions have not proven to be effective to improve the immune recovery, nor have they succeeded in reducing immune activation or levels of inflammatory markers in INRs, at least with the dosage schedule selected.