Immune Restoration With Antiretroviral Therapies: Title and subTitle BreakImplications for Clinical Management

Immune deficiency is the hallmark of human immunodeficiency virus (HIV) infection, and it places HIV-infected persons at risk for the opportunistic complications of acquired immunodeficiency syndrome (AIDS). The immune deficiencies of HIV infection are both quantitative and qualitative, and are characterized by a progressive decrease in circulating CD4 cells and functional impairment of the remaining CD4 cells, as well as defects in the function of other lymphocytes. Since the introduction of highly active antiretroviral therapy (HAART), there has been a dramatic decrease in HIV-related mortality. For example, at the University Hospitals of Cleveland John Carey Special Immunology Unit (Cleveland, Ohio), the annual observed number of deaths decreased by approximately 80% between 1995 and 1998,¹ reflecting national trends reported earlier.² These decreases in mortality are clearly related to availability of more effective antiretroviral drugs and an improved understanding of how to use them. A decreased rate of HIV replication results in significant improvement in both the number and function of circulating CD4 cells.^{3 - 5} The overall enhancement in immune capabilities results in a diminished risk of opportunistic infections (OIs) and death.

Immune restoration after treatment with HAART can be divided into 2 distinct phases with distinguishable kinetics and mechanisms. A first-phase rapid increase in circulating CD4 cells is observed within the first 8 weeks of therapy^{4 ,6 - 7} and is likely due largely to redistribution of lymphocytes from lymph nodes where circulating lymphocytes had been trapped by the inflammatory state characteristic of HIV replication.^{8 - 9} Decreased risk of OIs has been demonstrated in this early period.^{10 - 11} Thus, simple redistribution of lymphocytes from lymphoid tissues to the circulation may be sufficient or may collaborate with decreases in viral replication and virus-induced immune dysregulation to provide early protection from the major OIs that characterize HIV infection and AIDS.

Second-phase cellular restorations are more modest, characterized by a slower increase in CD4 cells, composed largely of naive cells, that likely reflects continued thymic production or, perhaps, extrathymic cellular division that is no longer counterbalanced by HIV-induced cellular sequestration and destruction. During this period, both CD4-naive cells and CD8-naive cells tend to increase while memory CD4 cells tend to stabilize or increase modestly and memory CD8 cells tend to decrease.^{7 ,12 - 14} Although the CD4 cell increase during this phase is modest, in most persons, the slow increase persists for at least the first 2 years of therapy.^{15 - 17}

Although the increase in circulating immune cells and improvements in the function of these cells are apparently sufficient to confer significant protection against OIs in most patients, among persons with moderately advanced or advanced disease, ongoing therapy for 1 year or more does not normalize all indexes of immune function. The CD4 cell counts are often below normal levels^{4 ,7} and the function of these cells can often be shown to be abnormal using a number of clinical indexes, such as skin testing for delayed-type hypersensitivity, or using laboratory assays such as those used to assess lymphocyte proliferation,⁷ cytokine production,¹⁸ or responses to immunization.¹⁹ Also, assays that are more sophisticated but of uncertain clinical significance, including analyses of T-cell receptor diversity,^{20 - 22} cell cycle progression²³ and cellular activation markers,^{5 ,24 - 25} may tend to show improvement yet remain abnormal despite long-term suppression of HIV replication. Whether prolonged suppression will normalize these indexes with time or whether these abnormalities reflect a heightened risk of morbidity for persons with HIV infection remains to be determined.

The answer to this question is clearly no. The factors that predict the magnitude of CD4 cell increase after initiation of HAART include the baseline level of HIV replication and the rate of CD4 cell decrease before the initiation of HAART.^{5 ,7 ,15} It is likely that these 2 indexes are related, since the bulk of the first year's increase in circulating CD4 cells is observed during the first-phase cellular redistribution period and this increase is due to the decrease in HIV-related inflammatory responses in lymphoid tissue.^{8 - 9 ,26} Another predictor of the greater magnitude of CD4 cell increase is a more gradual first-phase decay in plasma HIV RNA levels after initiation of HAART.²⁷ It is not clear why persons with a more rapid first-phase decay in plasma HIV RNA levels have a less prominent CD4 cell increase. We suspect that the turnover and death of HIV-infected cells during the first phase of decay and in HIV-infected lymph nodes may be related; ie, persons with a more rapid first-phase decay (greater cellular turnover) may have greater levels of cell death in lymphoid tissues, resulting in fewer viable cells to redistribute after HIV replication is blocked by antiretroviral therapy.

Thus, there is considerable variability in the magnitude of immune restoration following HAART. As mentioned, persons with higher baseline HIV RNA levels and more acute pretherapy CD4 cell declines have greater CD4 cell increases than persons with lower pretreatment HIV RNA levels and more subtle CD4 cell declines. In addition to these predictors, the CD4 cell increases described herein represent the mean or median changes in the group. The classic biphasic increase in circulating CD4 cell counts may not be experienced by all or even most persons receiving HAART.²⁸ A minority of persons may experience an actual decrease in circulating CD4 cells; on the whole, about 5% to 15% of persons who have excellent suppression of HIV replication after initiation of HAART experience no or only a modest increase in circulating CD4 cells.²⁹ Poor responses may be more frequent in HAART-treated patients with pretreatment CD4 cell counts of more than 500 × 10⁶/L; in 1 study, as many as 23% of such treated patients experienced a decline in circulating CD4 cells despite suppression of HIV replication.³⁰ It is not clear yet how much of this discordant decline is related to variability in lymphocyte enumeration or other artifacts of analysis. We examined a small group of patients who largely had more advanced disease (median pretreatment CD4 cell count of 160 × 10⁶/L) and poor CD4 cell increases despite excellent suppression of HIV replication for at least 1 year. These discordant responders tended to be older than persons who experienced greater CD4 cell increases and also tended to have higher CD4 cell counts before treatment than did patients in the same clinic with more robust CD4 cell increases. Preliminary findings suggest that thymic failure may underlie the discordant responses in these patients with more advanced disease,³¹ but the mechanisms of discordant responses in persons with higher baseline CD4 cell counts are unknown and the role of the thymus in those responses is unclear.

The optimal treatment of patients who experience successful control of HIV replication yet maintain low circulating CD4 cell counts is not clear. One option is to continue the existing treatment regimen, since some of these persons will in time experience a more substantial increase in circulating CD4 cells.¹⁵ Another option is to substitute another suppressive antiretroviral regimen, as there have been anecdotal reports of more successful cellular restoration after such revisions.³² Because interleukin 2 (IL-2) promotes the division of mature T cells, IL-2 administration may increase levels of circulating CD4 cells in this setting.^{33 - 35} The use of IL-2 for persons with immunologic failure despite virologic success has been approved for compassionate administration in France. Nonetheless, although IL-2 may increase circulating CD4 cell counts in this setting, the clinical benefit of CD4 cell increases after IL-2 administration remains to be proven and the clinical benefit of IL-2 therapy for HIV infection is currently being examined in 2 large international trials (SILCAAT [Chiron Corp] and ESPRIT [NIH and Chiron Corp]).

Another agent that is undergoing evaluation in persons experiencing therapeutic failures after administration of antiretroviral therapy is granulocyte-macrophage colony-stimulating factor, which increases circulating CD4 cell counts,^{36 - 37} perhaps via the increase in production or release of progenitor cells that can develop into mature T cells. Two newer agents that merit study in this setting include interleukin 7, which in animal models increases rates of thymic as well as extrathymic T-cell maturation,^{38 - 39} and interleukin 15, which in in vitro studies enhances lymphocyte proliferative responses and cytolytic lymphocyte activities in HIV and simian immunodeficiency virus infection.^{40 - 43} Because these 2 latter agents have not yet been used for treating humans, the utility of these strategies will not be tested for some time.

This is a critically important question, and the answer may help patients and physicians determine the best time to initiate antiretroviral therapy. Unfortunately, the data addressing this question are incomplete. It appears that treatment with HAART during or shortly after acquisition of HIV infection permits a greater restoration of immune function, including a preservation of HIV-specific CD4 cell responses^{44 - 45} that is rarely seen in persons with more advanced disease^{4 - 5 ,46} but may be occasionally observed in persons with earlier disease who have pretreatment CD4 cell counts of greater than 500 × 10⁶/L and who are treated with antiretroviral therapy.⁴⁷ Because CD4 cells may be needed to maintain cytolytic host defenses against HIV, preservation of these responses may provide HIV-infected persons a means to better control HIV replication in the presence (or, perhaps, even in the absence) of antiretroviral therapy. Although there are some indications that persons treated at this early point in infection may have better endogenous control of HIV replication⁴⁸ and, perhaps, a lower frequency of minor OIs in the short term,⁴⁹ few patients have been studied and the treatment regimen in the only controlled clinical end point trial was zidovudine alone. It is thus unclear if aggressive implementation of this strategy will affect the course of HIV disease. Thus, although most clinicians believe that antiretroviral therapy is indicated during or shortly after acute HIV infection, there are insufficient controlled data to be certain that this will result in a better long-term clinical outcome than if antiretroviral therapy is delayed.

In persons with chronic HIV infection, the relationship between the timing of treatment initiation and immune restoration also is not clear. On one hand, persons with higher pretreatment CD4 cell counts may be more likely to experience a paradoxical CD4 cell decrease during therapy.³⁰ On the other hand, the partial immune restoration observed after treatment with HAART may be more substantial in persons who are treated before peripheral CD4 cell depletion.⁴⁷ This stated, immunization after suppression of HIV replication can result in at least partial restoration of depleted CD4 cell responses to recall antigens as well as to HIV antigens.^{19 ,50} At present, however, we do not know if there are some aspects of the immunologic repertoire that are irretrievably lost after prolonged HIV replication. Lower CD4 cell nadirs are associated with both a greater risk of opportunistic complications after at least partial suppression of HIV replication⁵¹ and a greater risk of virologic failure.^{52 - 53} Controlled studies are needed to better understand the longitudinal effects of sustained HIV replication and the effects of HAART on restoration of immune responses when applied at different stages of HIV disease. The immunologic and virologic benefits of earlier treatment for HIV infection must be balanced with the toxic effects of combination antiretroviral therapy; this can best be accomplished by a randomized controlled trial.

As many as half of patients who begin treatment with combination antiretroviral therapy in the clinic setting tend to experience virologic failure as defined by plasma levels of HIV RNA of more than 400 copies/mL.^{52 - 53} Some of these persons have virus with high-level resistance to their therapy and, in many, sufficient levels of cross-resistance are present to render complete viral suppression unlikely using any available treatment regimen. Is continuation of therapy reasonable in this setting? Two small preliminary studies may provide some insight into this issue.

In a small pilot study, Miller et al⁵⁴ stopped antiretroviral therapy in patients who were infected with virus containing mutations conferring resistance to antiretroviral agents. After at least 8 weeks of treatment withdrawal, more than half of the patients experienced a reversion or outgrowth of plasma virus to wild type (ie, without evidence of drug resistance). This suggests that in the absence of drug, the presence of resistance mutations rendered that virus less fit in terms of replication than the wild-type virus. Levels of circulating CD4 cells tended to have a greater decrease in persons whose virus reverted to wild type than in persons with drug-resistant virus.

Deeks et al⁵⁵ extended these findings in a prospective virologic and immunologic analysis of a group of similar patients who underwent a supervised treatment interruption. Emergence of drug-sensitive virus was observed in almost all persons within 12 weeks of treatment interruption. Emergence of drug-sensitive virus was associated with a trend toward higher plasma HIV RNA levels and with a rapid decline in circulating CD4 cells. These studies have indicated in a more direct way that the concept of immunologic success in the presence of virologic failure is at least in part related to the diminished ability of drug-resistant virus to induce CD4 cell losses. Whether this is related to diminished replicative fitness of this virus or diminished ability of multidrug-resistant virus to induce cellular cytopathicity remains to be determined.

The clinical implications of these findings are that there appears to be some immunologic benefit to the continued selection pressure that is exerted by antiretroviral therapy even in the presence of resistance. So even if a treatment regimen is only partially or minimally suppressing HIV replication, continuation of treatment (if it is tolerated) seems reasonable. As a general rule, however, the best analyses to date suggest that the immunologic benefit of more complete virologic suppression is greater than the benefit seen in persons whose virus is incompletely suppressed.^{13 ,19 ,53 ,56} Thus, if a suppressive treatment regimen is available and tolerable, patients are likely to benefit from revision of the regimen.

This question is addressed in a number of recently completed and ongoing studies^{57 - 62} and summarized most recently by Kovacs and Masur.⁶³ Recommendations for management have been proposed by the Centers for Disease Control and Prevention.⁶⁴ As a general rule, for persons receiving primary prophylaxis for Pneumocystis carinii pneumonia, once the CD4 cell count has increased to more than 200 × 10⁶/L and has been stable at this level for at least 3 to 6 months, prophylaxis can be withdrawn safely. A recent retrospective study also demonstrated that the risk of Pneumocystis carinii pneumonia recurrence in persons whose levels of circulating CD4 cells have increased to at least 200 × 10⁶/L while receiving HAART is also very low (no cases reported among 246 patients followed up for 236 patient-years).⁶² For persons receiving primary prophylaxis for Mycobacterium avium complex (MAC) infection, a similar increase to more than 100 × 10⁶/L CD4 cells and maintenance at this level for at least 3 to 6 months should allow for safe withdrawal of primary prophylaxis. There have been anecdotal reports of successful withdrawal of secondary prophylaxis for prevention of MAC recurrence,⁶⁵ but data are insufficient to recommend this at present. Likewise, there have been reports of successful withdrawal of therapy to prevent recurrence of toxoplasmosis and cryptococcosis,^{66 - 67} but there are also insufficient data to help determine when or if secondary prophylaxis for cryptococcosis or toxoplasmosis can be safely withdrawn. In persons with cytomegalovirus (CMV) disease, there have been anecdotal reports of safe withdrawal of anti-CMV therapies after HAART-induced immune restoration, with prolonged duration of disease-free survival.^{68 - 71} Relapses may be observed after CD4 cell increases; patients with healed CMV retinitis should be examined carefully by an ophthalmologist before treatment is withdrawn and then monitored carefully with regular ophthalmologic evaluations.^{64 ,72 - 73}

Thus, although the circulating CD4 cell count can be used as a general indicator of risk of or short-term protection from OIs, the information provided by the CD4 cell count does not completely reflect the immunologic repertoire of any individual. In fact, OIs have been reported in persons with high CD4 cell counts, with apparent failure of specific CD4 cell responses to the responsible pathogens.^{72 - 73} Also, among HAART-treated patients, lower CD4 cell nadirs were associated with an increased risk of OIs, even in persons with CD4 cell counts higher than 200 × 10⁶/L.⁵¹ Nonetheless, for most persons, levels of circulating CD4 cells are a reasonable reflection of general immunocompetence.

Paradoxically, in some persons with subclinical CMV or MAC infection, a profound inflammatory flare may be observed that manifests as an inflammatory uveitis⁷¹ or vitritis in the instance of CMV^{74 - 75} or as a febrile inflammatory lymphadenitis in the instance of MAC.⁷⁶ These syndromes do not necessarily indicate failure of immune reconstitution but, rather, are morbid complications of increased levels of circulating immune cells and accessibility of immune cells to sites of infection, underscoring the concept that host defenses not only protect against infectious diseases but also contribute to the clinical manifestations of infection.

Clearly, even the incomplete immune restoration provided by HAART regimens has been enough to dramatically decrease morbidity and mortality due to classic OIs. How long this protection can last remains uncertain, but, to date, there has been no apparent increase in the occurrence of AIDS-defining OIs¹ in settings where HAART is available. It is not clear that the immune restoration observed with HAART will provide protection against malignant complications of HIV infection. Rates of Kaposi sarcoma occurrence began to decline before suppressive antiretroviral therapy became available^{58 ,77 - 80} ; this was at least in part likely related to the epidemiology of human herpesvirus type 8, which has been linked to this neoplasm. On the other hand, there has not yet been any consistent evidence that the rate of occurrence of HIV-related non-Hodgkin lymphoma has been affected by the availability of HAART.^{58 ,77 - 78 ,81} Conceivably, duration of immune suppression may be a major determinant of the risk for this neoplasm, and, if this is the case, there may be a delay between immune restoration and the effects of immune restoration on the occurrence of this malignancy. Alternatively, other mechanisms, such as B-lymphocyte activation, may contribute to the risk of B-cell lymphoma. Although hypergammaglobulinemia associated with HIV infection tends to normalize with administration of HAART,⁸² the effects of HIV infection on B-cell activation are neither fully characterized nor completely understood. Thus, at this time, it is not clear how much and what kind of immune restoration is needed to normalize risk for non-Hodgkin lymphoma or other malignancies or to normalize survival in persons with HIV infection. Likewise, it is not clear which indexes are the most critical determinants to study to understand how to monitor these risks.

There is reason to be concerned that the spectrum of morbidity and mortality in HIV disease is changing rapidly to include metabolic complications of therapies⁸³ and infectious complications, such as hepatitis C.⁸⁴ Of recent HIV-related deaths occurring in the John Carey Special Immunology Unit of University Hospitals of Cleveland (number of deaths ranging from 20 in 1998 to 32 in 1999), although OIs constituted less than 25% of deaths in 1999, end-organ failures constituted nearly half.¹ Importantly, the median CD4 cell count among the patients who died in our clinic has risen, from 0 × 10⁶/L in 1995 to 75 × 10⁶/L in 1999, and about 20% of recent deaths have occurred among patients with plasma HIV RNA levels below the limit of detection.¹ Thus, the underlying causes of mortality in HIV infection are changing and may be less related to uncontrolled HIV replication and immune deficiency than in the past.

In conclusion, suppression of HIV replication results in clinically significant restoration of immune response, which has decreased recent morbidity and mortality attributable to HIV infection. Immune restoration is often incomplete, however, and in some persons experiencing the poorest restoration, failure of thymic output may be responsible. At present, circulating CD4 cell count is the most available and useful index of immunocompetence and can be used to guide indications for prophylaxis and withdrawal of prophylaxis for certain OIs. Although there is reason to believe that earlier initiation of antiretroviral therapy may permit greater preservation of immune responses and clinical benefit, this has not been convincingly demonstrated, and the issue remains unresolved. Persons whose treatment regimen is failing may still experience immunologic and clinical benefit from continuation of antiretroviral therapy as the selection pressure for drug-resistant virus appears to result in the outgrowth of virus that may be less pathogenic. How much immune restoration is needed to normalize long-term survival for persons with HIV infection is not clear, nor is it clear that immune deficiency as measured by currently applied assays is the critical determinant of the occurrence of more recent AIDS-related mortality.