Efficacy of galactose and adalimumab in patients with resistant focal segmental glomerulosclerosis: report of the font clinical trial group

The requirements for enrollment into the FONT study have been described in full detail in a previous report [10]. The key inclusion criteria were: (1) primary FSGS confirmed by renal biopsy or by documented history of a disease-causing mutation [3]; (2) failure to respond to prior therapy with corticosteroids and at least one other immunosuppressive medication; (3) age 1–50 years at onset of proteinuria and 1–51 years at time of randomization; (4) estimated GFR ≥40 mL/min/1.73 m²; (5) Urine protein:creatinine ratio (Up/c) > 1.0 g/g on first morning void and (5) no immunosuppressive therapy for at least 30 days except low dose prednisone. Testing for podocyte gene mutations was not required prior to enrollment in the FONT II trial because renal fibrosis is an intrinsic feature of primary FSGS, regardless of whether or not there is a defined genetic mutation.

In this phase, subjects were placed on the maximal tolerated doses of lisinopril, losartan, and atorvastatin, based on reports of adverse effects and measurements of blood pressure, serum K⁺, creatinine, liver enzymes, and cholesterol concentrations. The doses of the angiotensin converting enzyme inhibitor (ACEi)/angiotensin receptor blocker (ARB) treatment had to be stable for a minimum of 2 weeks prior to randomization. No specific recommendations were made regarding daily dietary sodium or protein intake and this was left to the discretion of the site investigator.

All subjects received standard conservative medical therapy consisting of a combination of the following three agents: 1) lisinopril: 2) losartan; and 3) atorvastatin. The maximum target doses for subjects weighing <40 kg were: lisinopril 10 mg, losartan 25 mg, and atorvastatin 10 mg. For subjects weighing >40 kg, the maximal doses were: lisinopril 20 mg, losartan 50 mg, and atorvastatin 20 mg. The following novel therapies were administered for 6 months:

The dose of adalimumab was 24 mg/m² (maximum 40 mg/dose) every other week as a subcutaneous injection for the entire treatment period. Although the pharmacokinetics (PK) data from the FONT-I Study indicated enhanced clearance of adalimumab in subjects with FSGS and nephrotic-range proteinuria, the dose was not increased above the standard dosing regimen for other FDA approved indications to minimize the adverse event risk [14].

One subject was randomized to rosiglitazone treatment prior to this arm being dropped and replaced with galactose. This subject was not included in the analysis. The use of the full array of experimental therapies in the FONT Phase II Trial was authorized by the FDA (IND # 100,037).

The co-primary endpoints included:

The secondary endpoints were: (1) adverse effect profile; (2) percent change in proteinuria (evaluated as continuous variable); (3) change in eGFR, (4) time to halving of eGFR and ESKD

Figure 1 outlines the course of the Phase II trial. Subjects were evaluated after 2, 8, 16, and 26 weeks of the assigned experimental treatment. A follow-up evaluation was performed at 1 month, 3 months, and 6 months after discontinuation of the novel therapy, and then every 6 months until the end of the funding period. At each visit, blood was obtained for determination of CBC and a comprehensive metabolic profile and a first morning urine sample was obtained for measurement of Up/c. P_alb, a measurement of the capacity of serum to impair the filtration barrier to albumin, was determined at baseline in all subjects and after 8 and 26 weeks in subjects assigned to the galactose arm. The serum galactose concentration was also measured at baseline and after 26 weeks of treatment in those who were treated with galactose.

Stop points for novel therapy included: 1. >50 % decline from baseline eGFR and <60 mL/min/1.73 m² or a final level <20 mL/min/1.73 m²; 2. ESKD, i.e., initiation of dialysis or receipt of a renal transplant; 3. Serious adverse event (SAE), i.e. grade 4 CTC toxicity, version 4 (http://​evs.​nci.​nih.​gov/​ftp1/​CTCAE/​CTCAE_​4.​03_​2010-06-14_​QuickReference_​5×7.​pdf); 4. Increase in ALT/AST to >2.5×the upper limit of normal; 5. Onset of congestive heart failure or a myocardial infarction; 6. Clinical onset of SLE and/or positive ANA ≥ 1:160; 7. Serious infection including sepsis; 8. Malignancy; and 9. Pregnancy.

All routine serum and urine biochemical tests were performed in a central laboratory (Spectra East, Rockleigh, NJ).

Galactose was measured in the serum by GC/MS via modification of a published procedure [15]. In brief, 50 μL of 100 μg/mL U-13C galactose (Cambridge Isotope Labs) internal standard prepared in deionized water was added to 200 μL of serum. All other reagents were obtained from Sigma unless otherwise specified. In order to deplete glucose in the samples, 100 μL 2.5 M potassium bicarbonate, 50 μL 2.5 M phosphate buffer pH 6.5, 25 μL 300 U/μL glucose oxidase in 0.1 M potassium phosphate buffer pH 7 and 25 μL 50U/μL Catalase in 0.05 M phosphate buffer pH 7, were added and then incubated for 2 h at 37 °C with mixing. 200 μL each of 0.3 N zinc sulfate and 0.3 N barium hydroxide were added, the samples were vortexed and left on ice for 30 min. The samples were centrifuged at 3200 rpm for 10 min.

Prior to derivatization, the sample supernatants were deionized using cation exchange resin (AG50W-×8 hydrogen form) and anion exchange resin (AG 1-×8, formate form) (Biorad). Samples were eluted with deionized water and then dried. Galactose was derivatized to its acetate derivative for GC/MS analysis. 100 μL 20 mg/mL hydroxylamine hydrochloride/pyridine was added to the dried sample and heated at 70 °C for 30 min. 150 μL acetic anhydride was added with continued heat at 70 °C for another 30 min. The samples were transfer to vials with inserts for GC/MS analysis.

GC/MS analysis of galactose was done with an Agilent 6890 N Gas Chromatograph equipped with a 7683 Auto sampler and 5973 Mass Spectrometer and a DB 5 ms capillary column of 40 m × 0.18 mm × 0.18 μm (Agilent Technologies). Helium was used as carrier gas with a constant flow rate of 1.2 mL/min. Inlet temperature held at 280 °C. The initial oven temperature held at 180 °C for 30 sec and temperature ramps 50⁰/min to 220⁰ hold for 30 sec followed by 50⁰/min to 300 °C hold for 2 min. 1 μL of derivatized sample was injected in 5:1 split mode.

Mass spectrometer transfer line temperature held at 280 °C and operated in electron impact ionization mode at 70 eV. Selective Ion Monitoring (SIM) was used for fragments 314.1 and 319.1, 30 msec dwell time for each. Galactose concentrations were calculated from relative peak areas of ions 314.1/319.1.

P_alb was determined by the change in glomerular volume after exposure to media with a reduced albumin concentration and oncotic pressure in accord with previously published methods [16]. Serum suPAR was measured with the Quantikine Human suPAR Immunoassay (R&D Systems). Podocyte β3 integrin activity was determined by AP5 immunostaining of differentiated human podocytes in vitro, accord with previously published method [17].

Continuous data are reported as mean ± SD and in select instances the median (25, 75 percentile). As noted in the study design paper [10], when FONT II was initiated, the primary analysis of outcome was based on a hybrid Phase II design [18] with the study conducted in two phases. The first phase of the study would have ended when 17 participants had been randomized to each treatment arm. At that point, a ranking/selection comparison [19, 20] would have been done between the treatment groups and a minimum rate of success criterion would have been applied to each treatment group. After that, successful treatments would have moved forward to a second phase of investigation. An experimental therapy in which at least 2 out of the first 17 patients responded would have been advanced to the full Phase II assessment. A total of 42 subjects would have been assigned to each group (15 additional participants would have been enrolled in the galactose arm to ensure that 42 had a P_alb > 0.5). However, because of limited enrollment, the hybrid Phase II design could not be followed. Instead, the results were analyzed using the Freeman-Halton extension of Fisher’s exact test.

The study protocol, design and consent forms were approved by Institutional Review Board (IRB) at each participating site. This included the Institutional Review Board of North Shore/LIJ-Cohen Children’s Medical Center, Cincinnati Children’s Medical Center, Boston Children’s Hospital, Nationwide Children’s Hospital, Doernbecker Children’s Hospital, Carolinas Medical Center, University of Alberta-Stollery Children’s Hospital, Texas Tech University, Children’s Mercy Hospital, University of Kansas, University of Michigan, Columbia University-College of Physician’s and Surgeons, NYU School of Medicine-NYU Langone Medical Center, and University of Miami. Written informed consent was obtained from all participants or, where participants were children, from a parent or guardian. Assent was also obtained from children and adolescents in accord with guidelines at each site. The trial was authorized by the FDA under IND number 103,147. It was listed at ClinicalTrials.gov with the identifier NCT00814255 and the registration date was December 22, 2008.

Thirty-two subjects consented and enrolled in the study. However, only 21 subjects were randomized to one of the treatment arms (Fig. 2). The reasons that subjects withdrew or were dropped prior to randomization are detailed in the CONSORT diagram. The duration of the screening period was 1.8 (1.0, 2.3) (median (25, 75th percentile) months. Subjects were followed for 1.6 (1.2, 1.7) years after randomization.

The demographic and clinical features at baseline of the subjects who were randomized to one of the experimental treatments are summarized in Table 1.

The main clinical outcomes for the 21 participants are summarized in Table 2. When examining change in proteinuria alone, none of the subjects assigned to adalimumab therapy achieved a 50 % reduction in proteinuria. In contrast, 3 out of 7 subjects assigned to the galactose arm and 2 out of the 7 subjects assigned standard conservative treatment manifested at least a 50 % decline in proteinuria. The favorable response was sustained for 3–12 months after discontinuation of the galactose but maintenance of all other treatments. In contrast, the effect of standard conservative therapy was not sustained after the completion of the treatment period. The primary outcomes, based on the combined proteinuria and eGFR endpoint, are shown in Table 3. No significant difference in the rates of success was seen (P = 0.48). The change in Up/c over time in the 5 subjects who had a favorable response to the experimental treatment is illustrated in Fig. 3. There were no significant differences in clinical features (age, gender, race) or main laboratory data (eGFR, Up/c, serum albumin concentration) between the 5 subjects who achieved a 50 % reduction in proteinuria and the 16 other participants who had persistent proteinuria that was unresponsive to treatment (Additional file 1: Table S1).

The P_alb, suPAR, and AP5 values at the onset of the study and after 8 and 26 weeks of galactose treatment are depicted in Fig. 4. At the time of randomization, P_alb was ≥ 0.5 (normal P_alb: <0.2) in all 18 subjects (8 assigned to the galactose [1 child was withdrawn because of an elevated ANA titer in the baseline sample after randomization]), 4 to the adalimumab and 6 to the standard medical treatment arm) in whom the assay was performed. It declined in all 7 subjects who received oral galactose treatment and the value was within the normal range throughout the course of therapy (Fig. 4a). In contrast, the P_alb value was unchanged after the 26-week treatment period in 2 subjects who received adalimumab and 2 on standard medical treatment (data not shown). The decline in P_alb was significantly different in the galactose versus the two treatment groups (P = 0.003). The consistent decline in P_alb in the subjects assigned to galactose was associated with a 50 % or greater decrease in 3 of 7 subjects (see above). Plasma suPAR levels were initially elevated (>3,000 pg/ml) in 8 out of 22 (36 %) subjects who had measurement performed at baseline. There was no consistent change in circulating suPAR concentration in response to galactose therapy and no consistent relationship between suPAR levels and the change in proteinuria over the 26-week treatment period (Fig. 4b). Similarly, there was no consistent change in AP5 staining of podocytes incubated with subject serum during the course of any treatment and this biomarker did not correlate with the reduction in urinary protein excretion (Fig. 4c). There was no correlation between P_alb and suPAR or AP5 staining in the analyzed subjects.

The serious adverse events that occurred during the study are summarized in Table 4. The other adverse effects are itemized in Table 5. Although most of the subjects did experience at least one averse event during the treatment period, these side effects did not necessitate discontinuation of the assigned experimental treatment.

The only treatment that satisfied the criterion for success in this limited study of novel therapies for patients with resistant FSGS was galactose. Oral administration of this monosaccharide achieved at least a 50 % reduction in proteinuria in 3 out of 7 or 42 % of the treated subjects, 2 of whom maintained a stable eGFR. It is difficult to interpret the response to galactose therapy in response to the serial measurements of the circulating level of the sugar. Although there was no significant change in serum galactose concentration over the course of the study, this is not unexpected because galactose is rapidly metabolized to glucose. In addition, the serum galactose concentration was measured at arbitrary times and we did not standardize blood sampling in relationship to the oral dosing schedule (trough level or fixed interval after ingesting the monosaccharide). However, the consistent decrease in P_alb in all subjects who received galactose suggests that all of the subjects adhered to the assigned treatment and is consistent with prior experience with the sugar [12]. There was no evidence of adverse effects related to the use of galactose and future studies of this agent may benefit from a stable formulation that is already prepared ready for use. Our findings are somewhat at variance with the data of Sgambat et al. [21] who administered galactose to 7 pediatric subjects with steroid-resistant nephrotic syndrome (2 with post-transplant FSGS), none of whom achieved a reduction in proteinuria despite a fall in P_alb. Differences in age, confirmation of the diagnosis of FSGS, by renal biopsy, and inclusion of patients with recurrent disease may account for the modest differences between this negative report and our findings.

It is worth noting that while 0 out of 7 subjects who were assigned to adalimumab in the FONT trial achieved a favorable response to treatment, 4 out of 10 subjects who received adalimumab over 16 weeks in FONT I had at least a 50 % decline in proteinuria [14]. Preliminary data from the NEPTUNE cohort study indicate that there is genomic evidence of TNF activation in a subset of patients with primary FSGS (unpublished observations). It is plausible that future efforts to delineate a molecular signature of TNF activation may be useful in identifying individual patients with FSGS who are more likely to respond to a TNF antagonist. It is becoming increasingly apparent that primary FSGS is a heterogeneous disorder with some cases representing a genetic disorder of the podocyte and others arising from alterations in the integrity of the glomerular barrier due to a circulating factor (see below). It is anticipated that in the future, improved efforts to characterize patients based on their underlying genetic and molecular profiling will facilitate better selection of patients who have a greater likelihood of responding to a targeted therapy. Two out of the 7 subjects who were assigned to conservative medical therapy also achieved a 50 % reduction in proteinuria, which supports the need to include this control arm in any prospective trial of a novel therapeutic agent in patients with resistant FSGS.

There are several features about the FONT study design that are worthy of note. The entry criterion, Up/c > 1, is lower than a cut-off value of 2 used in most studies. As such, it might appear to be less stringent than the usual definition of partial response, namely a 50 % decline in proteinuria from an initial value exceeding 2. The definition used in this trial was an extension of the entry criterion into the FSGS Clinical trial [6]. In fact, the mean Up/c was 4.9 and exceeded 2 in all cases. Moreover, using a cut-off of Up/c > 2 did not alter the results. The study was open label because of a reluctance to implement sham injections in children. However, because the primary outcome involved objective laboratory measurements, it is unlikely that this influenced the outcomes of the trial.

Our data suggest that the permeability factors that are defined by the P_alb bioassay [22] and the currently available suPAR ELISA kits are distinct entities [17, 23]. While P_alb decreased with galactose administration in every subject after 8 weeks, suPAR levels were unchanged throughout the cohort regardless of treatment. In the analyzed samples, there was no correlation between these two potential biomarkers or between either marker and the proteinuria response. Moreover, neither measurement in isolation fully accounted for the persistence or decline in proteinuria in this study. This finding is in contrast to earlier studies which suggested that elevated suPAR levels may be a biomarker for primary FSGS [17, 23]. Recent data suggest that full length suPAR may cause proteinuria in the presence of specific antibodies, e.g. auto-antibodies against CD40, in patients who develop recurrent FSGS after receiving a kidney transplant [24]. A similar 2-hit process may be involved in the pathogenesis of proteinuria in response to other putative circulating factors. We urge caution in interpreting our data about P_alb and suPAR in this study because of the limited sample size and the collection of samples fairly late during the course of disease in these subjects with refractory FSGS. Further prospective studies are needed involving patients with all types of glomerular disease including FSGS to determine the prognostic implications of these circulating factors and their role in the pathogenesis of proteinuria and kidney injury.

It is recognized that nephrology ranks near the bottom among medical specialties in the performance and completion of clinical trials [25, 26]. Enrollment into the FSGS Clinical Trial that served as the antecedent to the FONT Trial was below expectations [27]. Poor subject comprehension of the adverse consequences of CKD, the adverse effects of many proposed treatments, and the high incidence of rapid decline in kidney function in patients with refractory proteinuria are among the explanations that have been offered to account for low enrollment into nephrology studies and specifically trials of glomerular diseases. It is likely that all of these factors contributed to the low recruitment into the FONT trial. Problems with subject recruitment underscore the need to create an infrastructure for clinical research and to design pragmatic trials that can be implemented in a timely manner to test potential treatments in patients with rare diseases like FSGS.

Two additional and specific factors may have played a role in limiting recruitment into the FONT Trial. First, by restricting the enrollment to patients who had failed steroids and at least one other immunosuppressive medication, we preselected for those with established disease of several years duration. These patients may be at very high risk of rapid deterioration in kidney function compared to patients evaluated soon after the onset of the disease. In fact, several of the treatment resistant subjects who were offered participation in the study experienced a rapid decline in GFR during the screening period and became ineligible. These observations suggest that it may be prudent to consider clinical trials of novel therapies earlier in the course of glomerular disease with persistent proteinuria. Such patients may have more reversible anatomic changes to glomeruli and the interstitium and their renal function may be more likely to remain stable during the course of the screening evaluation and experimental treatment period. Second, this trial transpired during a period when rosiglitazone and adalimumab were each receiving intense scrutiny because of reports of serious adverse events. Specifically, the FONT trial study was launched almost simultaneously with the publication of a meta-analysis that suggested that rosiglitazone increased the risk of cardiovascular complications in adults receiving the drug for the management of type 2 diabetes [11]. Although the FDA never questioned the rationale for testing thiazolidinediones in patients with resistant FSGS, the ongoing adverse publicity surrounding the use of PPAR-γ agonists dampened patient willingness to be randomized to that arm. As a consequence, we replaced rosiglitazone with galactose, a test agent with no known adverse effects. In addition, a report appeared at nearly the same time linking the adalimumab to the occurrence of serious infections and malignancy [28]. This diminished physician enthusiasm for adalimumab as a treatment arm in FONT even though the adverse consequences of adalimumab were documented in patients who receiving adalimumab in conjunction with other immunosuppressive drugs as treatment for rheumatological diseases. A follow-up report substantially downgraded the level of risk associated with adalimumab [29]. Finally, it is unclear what impact subject perception of a randomization scheme that compared an oral treatment thought to have minimal risk (galactose) to an injectable drug with a well-defined risk profile had on enrollment. Some potential participants expressed concern about a perceived imbalance in risk between the study arms and voiced a strong preference for galactose. They may have declined to participate in the trial without assurance that galactose was the test therapy they would receive. These observations remind us that clinical trials occur in a real world context and enrollment can be seriously impacted by a wide range of extraneous factors, some that are beyond the control of the investigative team and others that can be addressed during protocol development.