Biologic Therapies in the Treatment of Psoriasis: A Comprehensive Evidence-Based Basic Science and Clinical Review and a Practical Guide to Tuberculosis Monitoring
- First Online:
- Cite this article as:
- Sivamani, R.K., Goodarzi, H., Garcia, M.S. et al. Clinic Rev Allerg Immunol (2013) 44: 121. doi:10.1007/s12016-012-8301-7
- 2k Views
The treatment of psoriasis has undergone a revolution with the advent of biologic therapies including infliximab, etanercept, adalimumab, efalizumab, golimumab, certolizumab, alefacept, secukinumab, abatacept, and ustekinumab. These medications are designed to target specific components of the immune system and are a major technological advancement over traditional immunosuppressive medications. Herein, we present a comprehensive, unbiased comparison of these medications focusing on their differences. For example, TNF antagonists can differ in the way they are dissolved and administered, the effector molecules they can bind, serum peak and trough levels, the types of intracellular signals they can induce, the in vivo complexes that they can form, their protein structure, and their incidence and timing of rare adverse events, among other things. A critical review of the clinical studies that have tested the efficacy of these molecules is also presented including head-to-head comparison trials. The safety of biologics in terms of their long-term adverse events is discussed, as is their use in different types of psoriasis and in different patient populations. Finally, all anti-TNF agents have been associated with a variety of serious and “routine” opportunistic infections, particularly tuberculosis. For this reason, anti-tuberculosis testing both prior to the initiation of a biologic therapy and annually during treatment is pertinent. The uses and limitations of both the tuberculin skin test (TST) and QuantiFeron®-TB Gold (QFT) are discussed, as is the care of patients who present with latent tuberculosis infection prior to the initiation of biologic therapy. Recommendations for tuberculosis monitoring are provided.
Psoriasis is a chronic inflammatory skin condition that is estimated to affect approximately 2% of the general population . It presents in various forms, most commonly as chronic plaque psoriasis, also called psoriasis vulgaris. Patients with psoriasis vulgaris develop erythematous plaques that have well-defined borders and silvery “micaceous” scale. The Psoriasis Area and Severity Index (PASI) is a measure of disease severity in chronic plaque psoriasis . This scoring system is used in clinical trials where a PASI 75 response refers to a 75% or greater reduction in baseline PASI. Approximately 25% of patients with psoriasis will eventually develop psoriatic arthritis .
There is strong evidence in favor of psoriasis being an immune-mediated disease with T cells playing a central role. Although the exact autoantigen eliciting the pathogenic immune response is unknown, recent population-based genetic studies have confirmed the association between the MHC class I HLA risk allele Cw*0602 and psoriasis. Interestingly, these studies have also identified ERAP1 as another gene associated with psoriasis in Cw*0602-positive individuals. ERAP1 plays an important role in the processing of peptides for presentation by class I molecules such as HLA-C. Since ERAP1 variants only influence psoriasis susceptibility in Cw*0602-positive individuals , the extrapolation is that ERAP1 processes the self-antigen that is presented by the HLA-C molecule. Other loci associated with psoriasis also harbor genes involved in the immune response. These include IL12B, IL23R IL28RA, REL, IFIH1, TRAF3IP2, NFKBIA, and TYK2 .
Genetics aside, several other lines of evidence support T cells playing a central role in the pathophysiology of psoriasis [5, 6], particularly T helper type 1 (Th1), Th17, and CD8+ Tc17 cells , and IL-17-secreting γδ T cells [7–10]. The differentiation of Th1 cells is dependent on IL-12. These cells secrete interleukin (IL)-2, interferon (IFN)-γ, and tumor necrosis factor (TNF, formerly known as TNF-α). TNF is also secreted by a variety of other cells including keratinocytes and antigen-presenting cells. In contrast, Th17 cells and Tc17 cells produce IL-17, IL-22, and IL-26. Naive T cells are polarized to secrete Th17 cytokines in the presence of IL-1β, IL-6, IL-23, and TGF-β. Once differentiated, IL-23 is then required for their continued maintenance . Some Th17 cells also secrete IFN-γ, so there are no absolute rules that the T cells must follow. In the setting of psoriasis, cytokines including IL-12, IL-17, IL-21, IL-22, IL-23, TNF, and IFN-γ stimulate a cycle of inflammation that results in activation of dendritic cells and other antigen-presenting cells, proliferation and activation of keratinocytes [6, 12–18], and increased neutrophil chemotaxis . There is also a recent manuscript highlighting a role for IL-33 and its ability to augment the action of substance P .
Although standard therapies for psoriasis can be very effective, they can also be associated with significant toxicities and patients with psoriasis have expressed a high level of dissatisfaction with them [21, 22]. Over the past decade, new biologics (parenterally administered protein therapeutics) that specifically target TNF, the Th1 and Th17 pathway-inducing cytokines IL-12 and IL-23, and T-cell transmembrane proteins important for cell adhesion and co-stimulation have changed the landscape of treatment options for psoriasis and other autoimmune diseases. Herein, we review the clinical trials of these medications in psoriasis and the problems that might arise with respect to evaluating tuberculosis status in these patients. We also review in detail the mechanism of action of the biologic drugs, focusing on how they differ from one another.
Three Different Avenues Used by Biologics to Treat Psoriasis
Alefacept, a recombinant LFA-3-IgG1 fusion protein, was the first US Food and Drug Administration (FDA) sanctioned biologic agent for the treatment of psoriasis. It binds to CD2 on memory T cells and functions to block the CD2-LFA-3 co-stimulatory signal utilized by CD45RO+ memory effector T cells. Once bound to the CD2-expressing T cell, the Fc portion of alefacept can also be recognized by natural killer cells, targeting the memory T cell for killing. This may explain why some patients experience continued improvement after the medication is stopped and how remissions can last from months to years.
Overview of current biological therapies
Mechanism of action
Human fusion protein of first extracellular domain of LFA-3 fused Fc portion of human IgG1
LFA-3 portion binds to CD2 on memory T cells to block their activation. Fc portion binds to CD 16 on natural killer cells to induce apoptosis of memory T cells
Astellas Pharma US, Inc.
Chimeric (murine–human) antibody against TNF-α
Binds TNF to neutralize its effects
Centocor Ortho Biotech Inc.
Human fusion protein of the TNF receptor to Fc portion of IgG1
Binds TNF to neutralize its effects
Amgen® and Wyeth®
Human monoclonal antibody against TNF
Binds TNF to neutralize its effects
Human monoclonal antibody against TNF from human immunoglobulin transgenic mice
Binds TNF to neutralize its effects
Centocor Ortho Biotech Inc.
PEGylated recombinant, humanized antibody Fab′ fragment against TNF
Human monoclonal antibody against the p40 subunit of IL-12 and IL-23 from human immunoglobulin transgenic mice
Blocks the actions of IL-12 and IL-23
Centocor Ortho Biotech Inc.
Human monoclonal antibody against the p40 subunit of IL-12 and IL-23 isolated from human antibody phage display library
Blocks the actions of IL-12 and IL-23
Human fusion protein consisting of the extracellular domain of CTLA-4 linked to a modified Fc portion of human IgG1
Blocks activation of T cells by blocking the co-stimulatory signal from antigen-presenting cells by blocking binding of B7 protein on antigen-presenting cells
IV and SC
Human monoclonal antibody directed against IL-17A
Blocks the action of IL-17A
Novartis Pharma AG
TNF Antagonists: How they Differ
The biologics that target TNF do not have the exact same mechanism of action, as evidenced by the clinical phenomenon in which patients who have failed one TNF antagonist often respond to a second TNF antagonist [23–26]. Although it is common for patients who develop neutralizing antibodies to infliximab or adalimumab to respond well when switched to etanercept, the presence or absence of neutralizing antibodies to the first TNF antagonist plays only a minor role in predicting the response to a subsequent TNF antagonist [27, 28]. Clearly, the medications in this class were all designed to neutralize TNF, but how the TNF antagonists differ from one another has been a topic of ongoing investigation.
Structurally, infliximab is a mouse–human IgG1 chimeric anti-TNF monoclonal antibody. Adalimumab and golimumab are fully human anti-TNF monoclonal antibodies (IgG1). Etanercept is a fusion protein, described in detail below, and Certolizumab is a pegylated, humanized Fab′ fragment of an anti-TNF monoclonal antibody.
In terms of affinity for TNF, adalimumab, infliximab, and etanercept all have high affinities with Kd values in the sub-nanomolar range. Neutralizing TNF prevents it from interacting with its receptors, TNFR1 and TNFR2. TNFR1 is expressed by all nucleated cells, and TNFR2 is generally expressed only on endothelial cells and hematopoietic cells. sTNF has a higher affinity for TNFR1 and the transmembrane form of TNF (tmTNF) has a higher affinity for TNFR2. Binding of TNF to its receptor induces adaptor proteins to bind to the receptor’s cytoplasmic domain, inducing a signal transduction cascade. This leads to activation of nuclear factor kappa-B1 (NF-κB1). NF-κB1 is a rapid-acting transcription factor that regulates genes controlling a wide range of functions including cell proliferation, survival, and cytokine production. Factors secreted in response to TNF signaling include IL-1β, IL-6, and C-reactive protein (CRP), among others. NF-κB1 activation usually inhibits apoptosis by inducing production of the FADD-like IL-1β-converting enzyme (FLICE). Thus, neutralizing TNF will prevent proinflammatory cytokine secretion and cell proliferation. (However, depending on the cell’s state, it is possible for TNF to induce an alternative pathway favoring caspase-8 and caspase-3 mediated apoptosis.)
TNF is also expressed by TNF-producing cells as a transmembrane trimer (tmTNF). This trimer can be cleaved and released in an active soluble form by the TNF-α-converting enzyme (TACE), a member of the ADAM family of metalloproteinases (Fig. 1). TNF-producing monocytes/macrophages and T cells have low levels of TNF on their cell surface . The ability of TNF antagonists to bind to tmTNF may also affect the biological activity of these medications. By studying TNF-overexpressing cell lines, it has been determined that etanercept, adalimumab, and infliximab all bind well to tmTNF, but that the binding of infliximab and adalimumab is superior. This might be due to the fact that a single tmTNF molecule can bind up to three infliximab molecules, but only one etanercept molecule [29, 35]. The bivalent infliximab and adalimumab can also bind to multiple tmTNF molecules. Once bound, TNF antagonists will prevent tmTNF from interacting with TNFR1 and TNFR2. In addition to physically inhibiting tmTNF/TNFR interactions, the crosslinking ability of adalimumab and infliximab can induce a signal transduction cascade originating from tmTNF (reverse signaling). This, in turn, may lead to suppression of cytokine secretion, decreased T-cell proliferation, and apoptosis [36–39] (Fig. 1). Infliximab and adalimumab can also be directly cytotoxic to tmTNF-bearing cells by inducing antibody-dependent cellular cytotoxicity (ADCC) and complement-mediated cytotoxicity (CDC) (Fig. 1). Studies have also shown that once bound to TNF, both adalimumab and infliximab, but not etanercept, are able to bind strongly to FcγRII and FcγRIII receptors. Furthermore, etanercept is unable to bind to C1q  and adalimumab and infliximab induce ADCC much more potently than etanercept. Thus, differences in Fc receptor and complement C1q binding may contribute to the differences in efficacy of the TNF antagonists. This remains a topic of investigation, as some researchers have not observed the same results when human peripheral blood mononuclear cells were used . Although the mechanism might be debatable, the observation is clear: there is a loss of T cells and monocytes following infliximab therapy [42–44].
The vastly different dosing regimens and pharmacokinetics of the TNF antagonists also dramatically impact their biological effectiveness. This is due in part to the drugs’ route of administration, their half-lives, and their peak and trough serum concentrations. The bi-weekly dosing of etanercept results in serum concentrations with a very low peak-to-trough ratio. In contrast, the bolus IV dosing of infliximab results in huge variations in its serum concentrations. Depending on when the blood is drawn, the serum concentrations of infliximab can be 40-fold higher than the peak concentrations of etanercept [45, 46], and the peak concentration of infliximab can be more than 50-fold greater than its trough concentration. Importantly, clinical improvement with infliximab has been associated with higher trough concentrations . Compared to increasing the dose of infliximab, increasing the frequency of the drug’s administration should theoretically be a more effective means of raising the patient’s serum trough concentration.
Although there have been conflicting reports, it is evident that the TNF antagonists also reduce expression of other cytokines. Etanercept has been reported to reduce serum levels of IL-17 and IL-22 and to suppress IL-17 signaling [48, 49]. It was demonstrated in a clinical trial that etanercept reduced IL-23 produced by inflammatory dendritic cells and decreased downstream effector molecules including IL-17, IL-22, CC chemokine ligand 20, and beta defensin 4 . In contrast, infliximab has been reported to reduce serum levels of IL-15, and there have been mixed reports on infliximab’s effect on IL-17 production [50, 51]. Although this is an area in need of further investigation, it is becoming clear that the effects of TNF inhibition encompass other cytokines; yet another way in which the TNF antagonists can differ from one another.
Lastly, because etanercept is a TNF–receptor–IgG1 fusion protein, its specificity is much different than the anti-TNF monoclonal antibodies. While adalimumab and infliximab seem to be specific for TNF, etanercept also binds to other TNF receptor ligands such as the lymphotoxin (LT) family members, LTα3, LTα1β2, and LTα2β1 . Like TNF, LT has a transmembrane and a soluble form, and etanercept can bind to either [34, 53–55]. The biological significance of this increased binding spectrum is not known, but LT molecules play a role in autoimmune disease and are important members of the LT/TNF/LIGHT (LT-related inducible ligand that competes for glycoprotein D binding to herpesvirus entry mediator on T cells) network [54, 56]. Theoretically, this may represent a major deviation of etanercept’s mechanism of action from the other TNF antagonists.
Ustekinumab: Inhibition of Both IL-12 and IL-23
IL-12 is a heterodimer composed of subunits IL-12A p35 and IL-12B p40. The p40 subunit of IL-12 also combines with p19 to form a biologically active composite cytokine IL-23. The p19 subunit of IL-17 shows no biological activity by itself. IL-23 is important for the development and maintenance of Th17 cells . In mice, cutaneous administration of IL-23 can induce a psoriasis-like disease . Th17 and Th1 cells have both been implicated in psoriasis [7, 48]. Since ustekinumab is directed against the p40 subunit, it is able to inhibit the function of both IL-12 and IL-23. By inhibiting Th17 cell differentiation and survival, ustekinumab therapy should decrease IL-17 levels, which are known to correlate well with psoriasis severity . Neutrophils are one of the hallmarks seen on histologic examination of psoriatic skin. Interesting, IL-17 is also known to be critical for neutrophil recruitment , which might explain some of the therapeutic effect of ustekinumab in psoriasis.
Clinical Trials Showing the Effectiveness of Biologics in Treating Psoriasis
Clinical trials with either IV [59, 60] or IM  form of alefacept have been shown effective in the treatment of psoriasis. These trials were 12 weeks long with a primary end point at 2 weeks. Due to the dose response of the drug, patients receiving a higher dose of alefacept were more likely to achieve a PASI 75. Alefacept-treated patients had an overall 9% increased risk of adverse events . The most common adverse events noted in these studies were dizziness , nausea , infusion-related chills [59, 60], pharyngitis [60, 61], headache , and pruritus . Serious adverse events that were seen only in the alefacept treatment group included the presence of coronary artery disease in four subjects, cellulitis in three subjects, and myocardial infarction in three subjects out of a total of 537 patients. IV dosing was noted to have increased incidence of serious adverse effects compared to IM dosing .
Manufactured by Genentech, efalizumab was directed toward the CD11a subunit of LFA-1. It was voluntarily pulled from the market due to its associated risk of progressive multifocal leukoencephalopathy, an often fatal central nervous system infection.
Infliximab at either 3 mg/kg [63, 64] or 5 mg/kg [63–65] was shown to have an increased efficacy of reaching a PASI 75 at 10 weeks, in comparison to placebo. Efficacies were maintained over placebo for 46–50 weeks. The most common adverse events were rhinitis , transaminitis , sinusitis , and headache . Onset of action is rapid, with significant improvement seen within 2 weeks and maximum improvement by week 10.
Although all protein-based medication can induce the formation of anti-drug antibodies, murine–human chimeric proteins, like infliximab, are the most immunogenic . Interestingly, higher doses of infliximab are progressively less immunogenic  and concomitant use of infliximab with methotrexate seems to reduce the formation of anti-drug antibodies [67–69]; however, there is no randomized clinical trial data on the use of methotrexate in combination with infliximab in psoriasis. Clinically, a loss in therapeutic effectiveness is noted in subjects that develop anti-infliximab antibodies [63, 65]. There is also a correlation between the presence of anti-drug antibodies and the incidence of infusion reactions to infliximab. Specifically, concentrations of anti-infliximab antibodies at greater than or equal to 8.0 μg/ml predict a shorter duration of therapeutic effect (35 days vs. 71 days) and a higher risk of infusion reactions . This decreased therapeutic effect may be due to increased infliximab clearance via the formation of multivalent drug–anti-drug immune complexes.
Clinical studies with etanercept demonstrated it to be superior to placebo for the primary endpoint of achieving PASI 75 after 12 weeks of treatment [71–75]. A dose response was noted when comparing subcutaneous injections of 25 mg and 50 mg [72, 74]. A study in a pediatric population showed that after 12 weeks, subcutaneous dosing at 0.8 mg/kg showed an increased efficacy in achieving PASI 75 as compared to placebo . Onset of action is slower than other TNF antagonists, but patients can show continued improvement for up to 6 months. A small clinical trial investigated the safety and efficacy of introducing etanercept into the immunosuppressive regimen of patients already on methotrexate. The combined methotrexate–etanercept regimen increased the number of patients who were “clear or nearly clear” when compared to patients in which methotrexate was discontinued .
Upper respiratory tract infections, sinusitis, headaches, and injection site reactions were the most common side effects in adults [71, 72, 74, 75], whereas streptococcal pharyngitis and skin papillomas were increased in the pediatric study . Interestingly, the injection site reactions were more frequent in the first 12 weeks of therapy and approached placebo frequencies afterwards. Anti-etanercept antibodies developed during treatment, but the antibodies were not neutralizing [71, 72].
Adalimumab was found to have superior efficacy at the primary endpoint of achieving PASI 75 at either 12 or 16 weeks in comparison to placebo in several studies. One study showed a dose response when comparing placebo, 40 mg every week, and 40 mg every other week at the primary endpoint at 12 weeks . Two other studies found that 40 mg every other week was superior to placebo in achieving the primary endpoint of PASI 75 after 16 weeks of treatment [78, 79]. Onset of action is rapid, with significant improvement seen within 2 weeks and maximal responses seen at 12 to 16 weeks. The most common side effects were upper respiratory infections , nasopharyngitis , headache , and cellulitis . The presence of antibodies against adalimumab was correlated with a loss of response .
Golimumab was found to have superior efficacy at the secondary major endpoint of achieving PASI 75 at 14 weeks in comparison to placebo . The most common side effects included upper respiratory infections, nasopharyngitis, and headache. The presence of antibodies against golimumab did not correlate with a loss in response, and patients who were on methotrexate at baseline did not form anti-golimumab antibodies.
Certolizumab is a pegylated, humanized Fab′ fragment of an anti-TNF monoclonal antibody. Clinical trials of its use in psoriasis are currently ongoing.
Two studies demonstrated that subcutaneously administered ustekinumab at either 45 mg or 90 mg was superior to placebo for the primary endpoint of achieving a PASI 75 after 12 weeks [81, 82]. The initial 12-week period was followed by a crossover study in which ustekinumab was administered to the initial placebo group, resulting in similar treatment efficacies. The onset of action is evident by 2 weeks with maximum efficacy seen between weeks 20 and 24. The most common side effect found in these two studies was injection site reactions . Common adverse effects such as upper respiratory infections, nasopharyngitis, and headaches were not significantly different from the rates found in the placebo group.
Study subjects receiving ustekinumab could be categorized as full responders or partial responders, defined as those subjects that achieved PASI 50, but not PASI 75, by 28 weeks. Partial responders had increased prevalence of neutralizing antibodies against ustekinumab. By decreasing the frequency of administration to every 8 weeks, two thirds of the partial responders could be converted to responders .
Briakinumab has not received FDA approval as of yet, but has been studied in one clinical trial. The trial used progressively increasing doses and found no difference in response among different dosings of 200 mg every week for 4 weeks, 100 mg every other week, 200 mg every other week, or 200 mg weekly. Furthermore, all of the mentioned dosing regimens showed significantly increased efficacies over placebo . Similar to the ustekinumab trials, the primary endpoint of PASI 75 was achieved at 12 weeks. However, a follow-up report of the open-label extension from the same trial showed a dose response with some patients losing their PASI 75 status . The most common adverse events were injection site reactions, upper respiratory infections, nasopharyngitis, and headaches. One case of sepsis with peritonitis was noted in the briakinumab treatment group during the open-label extension .
Although the report on the open-label extension comments on the presence of antibodies against briakinumab in several treatment groups, there is no quantification of the prevalence of these antibodies or their role in treatment response .
Several other clinical trials with briakinumab and a summary of these studies are available through a press release on Abbot’s website . The press release contains details on four phase III studies where briakinumab is compared to placebo or methotrexate, and two trials where briakinumab is compared to etanercept. As of January 2011, Abbot has withdrawn briakinumab’s application for FDA and European Medicine Agency approval with plans to possibly submit applications for both in the future.
Biologics on the Horizon
Abatacept is a fully human fusion protein consisting of the extracellular domain of CTLA-4 linked to a modified Fc portion of human IgG1. This blocks activation of T cells by blocking the co-stimulatory signal from antigen-presenting cells. This drug has not received FDA approval for use in psoriasis or psoriatic arthritis, but a phase II study in psoriatic arthritis revealed that patients treated with 10 mg/kg of abatacept had significant improvement when compared to placebo . A phase I study investigating the use of abatacept for psoriasis has been completed , and further studies and results are expected. Interestingly, there are two reported cases of psoriasiform eruptions developing in patients after initiation of abatacept [88, 89].
Secukinumab is a fully human monoclonal antibody directed against interleukin 17A. This drug has not received FDA approval for use in psoriasis. Early clinical studies in patients with psoriasis have shown that secukinumab was more effective than placebo in achieving PASI 50 and PASI 75 within a 12-week period . A phase II trial in patients with plaque psoriasis has been completed and a phase III trial is currently ongoing . The details of the findings are awaited.
Other Forms of Psoriasis
There are many variants of psoriasis and related diseases. Publications concerning biologic treatments for these psoriasis variants and related diseases are limited to case reports and small pilot studies. For example, generalized pustular psoriasis is a serious but rare form of the disease in which TNF antagonists have been, in general, very successful [91, 92]. Acropustulosis (acropustulosis of Hallopeau or acrodermatitis continua) is a pustular eruption of the skin over the terminal phalanges that can be associated with destructive arthritis and nail plate disease. According to case reports, patients with acropustulosis respond well to TNF antagonists . Given that erythrodermic psoriasis often evolves from psoriasis vulgaris, it is not surprising that these patients also seem to respond to TNF antagonists [94–97]. In contrast, palmoplatar pustulosis is evidently refractory to therapy with TNF antagonists . Psoriatic arthritis has been reviewed elsewhere [99, 100].
Head-to-Head Comparison Trials
Few manuscripts to date have reported head to head comparisons between two or more different biologics. Two trials comparing briakinumab to etanercept showed that briakinumab was superior to etanercept in achieving PASI 75 at 12 weeks [101, 102]. One study comparing ustekinumab to etanercept for moderate-to-severe psoriasis demonstrated a superior therapeutic effect of ustekinumab over etanercept at achieving PASI 75 and PASI 90 . Although 27.9% of the patients in this trial had psoriatic arthritis, there was no mention of the trial drugs’ effectiveness in treating arthritis, and a health assessment questionnaire-disability index was not included in the manuscript. This is of particular importance because etanercept is especially effective in treating psoriatic arthritis. The trial was also designed to measure efficacy at 12 weeks, but etanercept takes 24 weeks to reach an optimal response with some patients showing continued improvement for 96 weeks [71, 74]. This is in contrast to ustekinumab, which usually shows its peak efficacy much earlier. Indeed, the story of the tortoise and the hare would have had a much different ending if the race was unfairly shortened.
Another study compared infliximab, etanercept, and adalimumab, focusing on their effectiveness in the treatment of psoriatic arthritis . Improvement in PASI scores was secondarily analyzed and the authors concluded that adalimumab- and infliximab-treated groups had lower PASI scores after 1 year. However, all groups showed a greater than 90% reduction in the PASI scores. Because the clinical differences are not significant and analysis of the PASI scores were a secondary endpoint, further studies with a primary focus on cutaneous psoriasis will be needed to better assess any differences between infliximab, etanercept, and adalimumab.
Another study compared adalimumab against methotrexate. At 16 weeks, the primary endpoint (PASI 75) achieved by subjects in the methotrexate and the adalimumab treatment groups were 35.5% and 79.6%, respectively . However, the methotrexate was started at a low dosage and increased slowly over time, so 16 weeks may have been too short to appropriately evaluate the methotrexate response. After subgroup analysis was performed, it was suggested that methotrexate should be evaluated for 12 weeks and poor responders should be switched away from methotrexate . However, previous studies have shown that patients with psoriasis have variable methotrexate absorption, ranging as low as 32% . Thus, in a head-to-head clinical trial the actual absorption of methotrexate should be measured to identify the subjects that absorb methotrexate poorly and insure that these subjects are not underdosed. A subgroup analysis from this same parent trial used a composite score that combines assessment of response along with the presence of adverse events. This study reported that adalimumab had a better adverse event free response compared to methotrexate . However, we find it odd that the adverse event free response of adalimumab was also statistically superior to placebo. Evidently, patients who were able to reach a PASI 75 without treatment were more likely to experience an adverse event than patients reaching a PASI 75 on adalimumab. Because the methotrexate was underdosed in this short trial over 16 weeks, the composite score is skewed toward interventions that generate a faster response, and it is no surprise that adalimumab had an advantage over underdosed methotrexate. Interestingly, the adverse free response measure showed no difference between methotrexate and placebo, further supporting that the methotrexate was underdosed. This manuscript also lacked some important details. For example, adverse events with missing end dates were assumed to continue for the duration of the study, but it is unknown whether the prevalence of missing dates was similar between the different study groups. Also, the parent study reported PASI 50, 75, and 90 but the subgroup analysis calculated adverse event free days for only patients reaching PASI 75.
Briakinumab was compared to methotrexate as well, and the primary endpoint of the achievement of PASI 75 at 24 weeks was 81.8% and 39.9% for briakinumab and methotrexate, respectively . The full manuscripts detailing these studies are awaited.
Clinical Trials are not able to Assess Low Frequency Long-Term Adverse Events
An important concept to understand is that trials are powered to establish efficacy, not adverse events. Thus, rare but severe adverse events will likely not be detected in clinical trials, especially if the adverse event takes time to surface, longer than the duration of a typical trial . In addition, for ethical reasons, clinical trials in psoriasis usually maintain their placebo arm for only 3 months. After this point, the trial either finishes or the placebo group crosses over into a treatment arm(s). We are therefore dependent on limited registry data and reports to the FDA for monitoring of serious adverse events.
For example, it took analysis of reports to the FDA to demonstrate an association between TNF antagonists and heart failure . Also helpful in establishing this causal relationship was the fact that a clinical trial of etanercept for the treatment of severe heart failure had to be stopped early because there was an increased mortality in the etanercept arm . There are now numerous reports to the FDA demonstrating that even healthy individuals can develop heart failure on TNF antagonists. However, the initial clinical trials of these medications in rheumatoid arthritis (RA) and psoriasis failed to identify this risk because of its low incidence and the fact that the average onset of heart failure was 9.8 months for etancercept . The average onset of heart failure in patients treated with infliximab was shorter, around 3 months. Similarly, the association of neurological disease with TNF antagonists has largely come from case reports, registry, and FDA data . The American Academy of Dermatology recommends that TNF antagonists be avoided in patients with a personal history or a first-degree relative with multiple sclerosis or other demyelinating disorder [112, 113].
There have also been reports of granulomatous reactions, vasculitis, uveitis, and even psoriasis developing in patients on TNF antagonists [114–117]. These adverse reactions usually, but not always, develop in patients on TNF antagonists for rheumatologic indications. It is uncertain why these reactions occur, but we hypothesize that patient variation in the processing of biologic–TNF complexes may be responsible for some of the cases. Studies have also demonstrated that chronic exposure to TNF can downregulate TCR signaling and impair activation of T cells [118–120], a process that can be reversed with anti-TNF therapy [118–120]. In addition, there is a physiologic role of TNF in the development and function of follicular dendritic cell networks and germinal centers, important for humoral responses [118–120]. In the setting of molecular mimicry [121, 122], it is conceivable that TNF withdrawal may result in activation of autoreactive T cells. Antinuclear antibodies have also been reported to develop in patients treated with TNF antagonists .
Another complicating factor is that the genetic mutations that predispose an individual to autoimmunity can also predispose them to the development of cancer . Thus, it can be extremely difficult to tease out long-term cancer risks for the biologics. Rare associations between infliximab and the development of hepatosplenic T-cell lymphoma have been reported . In addition, there have been reports of early onset lymphoma after initiation of TNF antagonist therapy with improvement following discontinuation of therapy [126, 127]. A large meta-analysis of RA patients treated with infliximab or adalimumab demonstrated that patients on TNF antagonists have an increased risk of lymphoma . However, this study suffered from the problems described previously. Specifically, there was inclusion of open-label extension data without the comparable placebo arm, and there was no adjustment for the duration of exposure to the biologic therapy. In summary, the association between lymphoma and biologic therapy remains a highly controversial topic.
There does seem to be an increased risk of nonmelanoma and even melanoma skin cancer in patients with RA on biologic therapies . This is not surprising, given the role of the immune system in preventing UV-induced skin cancer . Although we are unaware of similarly robust studies conducted for psoriasis, cancer risk remains an important topic to discuss with your patients . With the exception of lymphoma and skin cancer, there is no evidence to support an increase in solid tumors with TNF antagonists.
In terms of pregnancy, it is difficult to assess the risk to the fetus given that there are often several potential contributing factors in autoimmune patients receiving TNF antagonists. Nonetheless, it is clear that infliximab passes through the placenta as high levels can be detected for extended periods in the serum of infants born to mothers on infliximab therapy . Infliximab has not been found in breast milk, however. There are several reports of women on TNF antagonists giving birth to healthy babies with no complications [132–134], but analysis of the FDA database revealed 61 congenital abnormalities in 41 children born to mothers on TNF antagonists .
The Role of Anti-Tuberculous Testing
TNF, IFN-γ, IL-12, IL-23, and IL-17 all play important roles in the immune response and, subsequently, control of Mycobacterium tuberculosis (TB) infection [136–140]. Mouse models have demonstrated that TNF is required for a protective immune response against TB , especially with respect to granuloma formation and inhibition of bacterial dissemination [142, 143]. Since the new biologics disrupt the function of TNF and other important cytokines, they will predispose patients to TB infection. This might seem obvious, but the link between TNF antagonists and TB infection did not take center stage until infliximab had already been administered to approximately 147,000 patients . Analysis of reports to the FDA revealed 70 reported cases of TB in patients treated with infliximab. The median onset was 12 weeks . Data from the Spanish Society of Rheumatology Database on Biologic Products (BIOBADASER) reported that there were 1,893 cases of TB per 100,000 patient-years of infliximab . Following these studies, there has been more careful selection and monitoring of patients, which has reduced the incidence of TB in patients receiving biologic therapy. Of the TNF antagonists, etanercept has been reported to have the lowest incidence of tuberculosis [146, 147] and has a longer lag time to reactivation of latent TB .
Suggestions for tuberculosis screening
– Not as sensitive or specific as QTF
– Best tested before start of methotrexate or biologics
– More widely available
– Possibility of pathergy to cause a false positive
– Requires a follow-up exam after placement within 48–72 h
– False positive with BCG vaccination
– More sensitive and specific than TST
– Best tested before the initiation of biologics
– Requires a blood draw and less widely available
– May lead to false negative during ongoing biologic therapy. Use in conjunction with exposure/travel history and symptomology
– False positive with exposure to Mycobacterium marinum
– Annual testing
Exposure and travel history
– Travel to endemic areas or exposure to known TB contacts elevate risk for
– Especially important after initiation of biologics as positive exposure or travel history should prompt a chest X-ray, even with negative TST or QFT
– Exposure to radiation
– Use in those with history concerning for possible TB exposure as TST and QTF are less reliable in ongoing biologic therapy
– Can follow internal lung changes in time
– Consider annual testing in those that are already on biologic therapy
For patients who have been determined to have LTBI, it is preferred that they complete the full prophylactic course prior to the initiation of biological therapy [139, 160, 161]. Starting biological therapy before the completion of the full prophylactic course is debated. The National Psoriasis Foundation recommends that at least 1–2 months of prophylaxis be completed prior to initiation of biological therapy . This recommendation is based on four references to expert opinions. In one reference, the authors suggest 1–2 months of chemoprophylaxis  and in another reference, the French guidelines suggest 3 weeks of a two-drug chemoprophylaxis regimen with one of the drugs being pyrazinamide . Of note, two of the references do not offer any recommendations regarding chemoprophylaxis [160, 161] prior to starting biological therapy. On the other hand, in one of the references, a case is reported of a patient who ultimately developed TB despite receiving chemoprophylaxis for 2 months prior to initiating anti-TNF therapy .
There are no primary studies to refute or support any of these recommendations, and the clinical urgency of starting biological therapy must be balanced against “sufficient” chemoprophylaxis for LTBI. It is known that full completion of a course of isoniazid decreases the chance for development of full TB by approximately 70–80% but not by 100% . In addition, the clinical presentation of TB infection in the setting of biologic therapy is often atypical, with at least 50% of the cases being extrapulmonary [144, 163]. The onset of clinically evident disease can be late, an average of 11.5 months for etanercept, 5 months for adalimumab, and 3 months for infliximab [144, 163]. Additional immunosuppressive medications, delayed diagnosis, and disseminated disease may contribute to the higher associated rates of morbidity and mortality for patients on TNF antagonists [144, 163]. Therefore, it must be emphasized to patients that there will be a possible risk for the development of TB even if the TB screening is negative or if the patient undergoes treatment for LTBI [160, 164, 165].
The development of biological therapies has led to several new effective options for the treatment of chronic plaque psoriasis, the most prevalent form of psoriasis. The following generalizations can be made from literature regarding the biological treatment of psoriasis.
Few studies have incorporated methotrexate. One study compared adalimumab against methotrexate  and only one published large randomized clinical trial has compared TNF antagonists head-to-head . In the study that compared adalimumab to methotrexate, the latter was started slowly with a low dose in the beginning. This may have contributed to the lower PASI 75 seen at 16 weeks . Future studies will need to escalate the methotrexate more quickly to mimic how the drug is usually dosed clinically. The variability of methotrexate absorption in the intestines of psoriatic patients  needs to be taken into account when interpreting possible “methotrexate treatment failures” in the literature.
The formation of antibodies against the biologic is not uncommon and can affect the long-term efficacy of the drug. Studies from the rheumatology literature show that combined dosing of a biologic agent with another immunosuppressive agent, such as methotrexate, decreases the formation of antibodies against the biologic agent [166, 167].
Resistance to one biological drug does not imply resistance to another. However, it is inconvenient to continually switch agents given the chronic nature of psoriasis. A better solution may be to treat patients with methotrexate along with the biologics. Although many clinical trials in the rheumatological literature have investigated the role of concurrent therapy with biologics and methotrexate, only a few case reports and an open-label pilot clinical study have detailed the utility in combining methotrexate with a biologic for the treatment of psoriasis [76, 168]. More studies with methotrexate are needed to better understand both the comparative efficacy against the biologics and its role in concurrent therapy.
Summary of biologics in clinical trials for psoriasis
Efficacy at primary or secondary endpoint
Antibody formation against biologic
A. At 2 weeks after treatment phase, reduction in mean PASI (primary endpoint) was 21%, 38%, 53%, and 53% in the placebo, 0.025 mg/kg, 0.075 mg/kg, and 0.15 mg/kg treatment groups, respectively. Patients achieving 75% reduction in PASI were 10%, 21%, 33%, and 31% in the placebo, 0.025 mg/kg, 0.075 mg/kg, and 0.15 mg/kg treatment groups, respectively. 
A. One patient developed “low” antibody titer
A. Higher dropout rate in the placebo group; data collection and analysis was performed by employees at sponsoring company
B. At 2 weeks after first treatment phase, a 75% reduction in the PASI (primary endpoint) was 4% and 14% in the placebo and the 7.5 mg treatment groups, respectively 
B. 5 patients developed “low” antibody titers
B. Higher dropout rate in the placebo group
C. At 2 weeks after treatment phase, reduction in mean PASI (primary endpoint) was 21%, 34%, and 44% in the placebo, 10 mg, and 15 mg treatment groups, respectively 
C. 4% of patients tested in alefacept-treated patients; antibodies were not neutralizing and had titers <1:40; one of the placebo patients had anti-alefacept antibodies
C. Study was underpowered at the primary endpoint of 15% mean reduction of PASI scores at 2 weeks after treatment for 10 mg treatment group; higher dropout rate in the placebo group; data analysis performed by study sponsor
A. At 10 weeks, a 75% reduction in the PASI (primary endpoint) was 6%, 72%, and 88% in the placebo, 3 mg/kg, and 5 mg/kg treatment groups, respectively 
A. 27% and 20% of patients in 3 mg/kg and 5 mg/kg, respectively
A. Power analysis not reported; higher dropout rate in the placebo group; site of data analysis not specified
B. At 10 weeks, a 75% reduction in the PASI (primary endpoint) was 3% and 80% in the placebo and infliximab treatment groups, respectively 
B. Cumulatively 27% of patients formed antibodies; antibody formation associated with loss of response
B. Data analysis performed by study sponsor
C. At 10 weeks, a 75% reduction in the PASI (primary endpoint) was 1.9%, 70.3%, and 75.5% in the placebo, 3 mg/kg, and 5 mg/kg treatment groups, respectively. Higher response efficacies were noted in the scheduled treatment group in comparison to the “as needed” treatment group 
C. At week 66, 49% and 39% of patients formed antibodies in the 3 mg/kg and 5 mg/kg treatment groups, respectively; 61.5% of titers were <1:40; antibody formation was related to loss of response
C. Higher dropout rate in the placebo group; site of data analysis not specified
A. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 2% and 30% in the placebo and etanercept treatment groups, respectively. At 24 weeks, a 75% reduction in the PASI was 5% and 56% in the placebo and etanercept treatment groups, respectively. At 24 weeks, DLQI improvement was 7% and 65% in the placebo and the 25 mg treatment groups, respectively 
A. Not reported
A. Higher dropout rate in the placebo group; site of data analysis not reported
B. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 4%, 14%, 34%, and 49% in the placebo, low, medium, and high treatment groups, respectively. At 24 weeks, a 75% reduction in the PASI was 25%, 44%, and 59% in the low, medium, and high treatment groups, respectively. At 24 weeks, DLQI improvement was 7% and 65% in the placebo and the 25 mg treatment groups, respectively 
B. Eight patients developed antibodies and no titers reported
B. Not sufficient power to detect difference between placebo and low treatment group; data analysis was performed by the sponsor
C. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 3%, 34%, and 49% in the placebo, 25 mg, and 50 mg biweekly treatment groups, respectively 
C. 1.1% and 1.6% developed antibodies in first and second treatment phases, respectively; antibodies did not affect efficacy; 73% of these patients had no antibodies at subsequent testing
C. Retrospective power analysis; higher dropout rate in the placebo group; data analysis performed by sponsor
D. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 5% and 47% in the placebo and the etanercept treatment groups, respectively. During the open-label period, PASI 75 levels decreased with duration of therapy 
D. 18.3% of patients had antibodies and titers were not reported; presence of antibody did not affect efficacy of treatment
D. Higher dropout rate in the placebo group; data analysis performed by investigators; tachyphylaxis with duration of therapy although this was related to presence of antibodies
E. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 11% and 57% in the placebo and the etanercept treatment groups, respectively. Placebo group approached PASI levels of treatment group during open-label treatment phase. Withdrawal–retreatment phase data was not reported 
E. Not reported
E. Retrospective power analysis; data storage and analysis performed by the sponsor; higher placebo dropout rate; treatment group had higher rate of streptococcal pharyngitis
A. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 4%, 53%, and 80% in the placebo, 40 mg every other week, and 40 mg weekly treatment groups, respectively. Efficacies of achieving PASI 75 decreased with duration of therapy 
A. Not reported
A. Efficacy of achieving PASI 75 decreased with duration of therapy; site of data analysis not reported
B. At 16 weeks, a 75% reduction in the PASI (primary endpoint) was 18.9%, 35.5%, and 79.6% in the placebo, methotrexate, and adalimumab treatment groups, respectively 
B. Not reported
B. Data analysis was performed by sponsor; higher placebo dropout rate; methotrexate started low with slow increase of dosage
C. At 16 weeks, a 75% reduction in the PASI (primary endpoint) was 7% and 71% in the placebo and adalimumab treatment groups, respectively. All patients that achieved PASI 75 at week 16 had a 92% improvement in their PASI by week 33. Re-randomization to placebo in withdrawal phase led to higher loss of response 
C. 8.8% of adalimumab-treated patients developed antibodies at some point during the study; titers not reported; presence of antibody correlated with loss of response
C. Data analysis performed by sponsor; higher placebo dropout rate
A. At 16 weeks, a 75% reduction in the PASI (secondary endpoint) was 2.5%, 40%, and 58% in the placebo, golimumab 50 mg, and golimumab 100 mg treatment groups respectively. At week 24, the 75% reduction in the PASI was 1%, 56%, and 66% in the placebo, golimumab 50 mg, and golimumab 100 mg treatment groups, respectively 
A. 4% of golimumab-treated patients developed antibodies with titers that ranged to 1:2,560; presence of antibody was not correlated to response; no antibodies formed in patients on methotrexate
A. Data analysis site not reported; higher placebo dropout rate
A. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 3.1%, 67.1%, and 66.4% in the placebo, 45 mg, and 90 mg treatment groups, respectively. By week 40, placebo crossover groups had similar efficacies to ustekinumab treatment groups 
A. 5.1% developed antibodies with titers that were <1:360
A. Data analysis performed by sponsor; higher dropout rate in the placebo group
B. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 3.7%, 66.7%, and 75.7% in the placebo, 45 mg, and 90 mg treatment groups, respectively. Partial responders did not benefit from escalated dosing at 45 mg but had higher PASI 75 rates with escalated dosing at 90 mg 
B. At week 52, 12.7% and 2% of partial responders and full responders had antibodies, respectively; the antibodies were neutralizing
B. Data analysis performed by sponsor; higher dropout rate in the placebo group
A. At 12 weeks, a 75% reduction in the PASI (primary endpoint) was 3%, 63%, 90%, 93%, 93%, and 90% in the a, b, c, d, e, and f treatment groups, respectively 
A. Not reported
A. Data analysis was performed by sponsor and investigators; high placebo dropout rate
The studies reviewed here have focused on the therapy of chronic plaque psoriasis, as this is the most prevalent form of psoriasis. Small reports have suggested that some of the biologics may be useful in treatment of other forms of psoriasis as well [169–171].
The development of biological therapies has revolutionized psoriasis treatment and many more biological therapies are on the horizon, including the targeting of IL-17. The research of biological therapies is starting to rapidly expand making this an exciting time for patients and physicians alike. As more studies are performed, more clinical experience is gained, and more long-term side effects and efficacies are identified, we will be able to determine which drugs are the most suitable for the long-term treatment of psoriasis.
EM holds career awards from the Burroughs Wellcome Fund and the Howard Hughes Medical Institute. The authors would like to thank Stephanie Chu for help with design and creation of figures and Ern Loh for providing the clinical photo.