Long term Tolerability and Clinical Outcomes associated with Tocilizumab in the Treatment of Refractory Antibody Mediated Rejection (AMR) in Pediatric Renal Transplant Recipients

Abstract

Treatment options for antibody-mediated rejection (AMR) are limited. Recent studies have shown that inhibition of interleukin-6 (IL-6) /interleukin-6 receptor (IL-6R) signaling can reduce inflammation and slow AMR progression. We report our experience using monthly tocilizumab (anti-IL6R) in 25 pediatric renal transplant recipients with AMR, refractory to IVIg/Rituximab. From Jan 2013 to June 2019, a median (IQR) of 12 (6.0–19.0) doses of tocilizumab were given per patient. Serial assessments of renal function, biopsy findings, and HLA DSA (by immunodominant HLA DSA (iDSA) and relative intensity score (RIS)) were performed. Median (IQR) time from transplant to AMR was 41.4 (24.3–67.7) months, and time from AMR to first tocilizumab was 10.6 (8.3–17.6) months. At median (IQR) follow up of 15.8 (8.4–35.7) months post-tocilizumab initiation, renal function was stable except for 1 allograft loss. There was no significant decrease in iDSA or RIS. Follow up biopsies showed reduction in peritubular capillaritis (p=0.015) and C4d scoring (p=0.009). The most frequent adverse events were cytopenias. Tocilizumab in pediatric patients with refractory AMR was well tolerated and appeared to stabilize renal function. The utility of tocilizumab in the treatment of AMR in this population should be further explored.

1. Introduction

Pediatric kidney transplantation is the preferred treatment for children who require renal replacement therapy since the 1960s, when the first pediatric kidney transplant was performed¹. Antibody mediated rejection (AMR) causes significant morbidity in both pediatric and adult kidney transplants recipients and remains a primary cause of progressive allograft dysfunction and failure^2,3. AMR is defined by clinicopathological features, which include chronic vascular changes in the form of glomerulopathy, multilamination of the peritubular capillary basement membrane, and arteriopathy⁴. Donor specific antibodies (DSA) to human leukocyte antigen (HLA) class I and II as well as antibodies to non-HLA targets⁵ are central in the pathogenesis of AMR.

Although significant advances in the treatment of acute cellular rejection (ACR) have been accomplished, AMR remains a challenge⁶. Available treatments for AMR include plasmapheresis, intravenous immunoglobuin (IVIg), pulse steroids, B-cell depleting antibodies, and proteosome inhibitors. These treatments focus on the reduction of HLA DSA, which has been shown to improve allograft outcome^7–9. Despite the availability of these options, there is no standardized approach to treating AMR and there remains no reliable way to predict treatment response. Furthermore, many patients will not respond to these therapeutic interventions, leaving them at significant risk of early allograft failure^10,11. The data in the pediatric kidney transplant population regarding chronic AMR is limited. However, it is known that pediatric patients develop de novo DSAs more frequently than adult kidney transplant patients¹² and this can precipitate the development of chronic AMR¹³. Similar to adults, AMR has been associated with a poor prognosis in pediatric kidney transplantation¹⁴, but it is especially problematic in the pediatric population where patients are expected to require multiple renal transplants during their lifetime

Tocilizumab (Actemra, Roche/Genentech, San Francisco, CA) is a humanized interleukin-6 (IL-6) receptor monoclonal antibody that is FDA approved for the treatment of various autoimmune rheumatological disorders and cytokine release syndrome. IL-6 is a proinflammatory cytokine that is critical in regulation of the differentiation of CD4+ cells, B cells and plasma cells. IL-6 levels are known to be elevated in the blood and urine of kidney transplant recipients during acute rejection episodes and have been shown to normalize after treatment^15,16. More recently Choi et al described stabilization of renal function and reduction in HLA DSA in highly sensitized patients with chronic active AMR who were unresponsive to IVIg, rituximab and plasmapheresis who received tocilizumab therapy¹⁷. The effects of tocilizumab in pediatric kidney transplant recipients with AMR have not previously been reported. This is the first report describing the tolerability and clinical outcomes of pediatric kidney transplant recipients who received tocilizumab as rescue therapy for refractory AMR.

2. Materials and Methods

2.1. Patient Selection and Treatment Protocols:

This is a dual center retrospective study that was approved by both institutions (IRB protocol Pro 00057757 and #21–000487 for Cedars-Sinai Medical Center and University of California, Los Angeles (UCLA) respectively). We identified 25 patients (13 from UCLA and 12 from Cedars-Sinai) who had evidence of active or chronic active AMR on histology between January 2013 and June 2019 and were refractory to treatment with IVIg +/− rituximab, pulse steroids, plasmapheresis and/or bortezomib (Table 1). The dose of IVIG was 1–2g/kg/dose, the dose of rituximab was 375 mg/m², and the bortezomib protocol was administered as previously described¹⁸. All except 2 patients were > 12 months post-transplantation at the time of initial AMR diagnosis. Patients were considered refractory to treatment if repeat biopsy/biopsies after initial diagnosis showed continued active or chronic active AMR despite other therapies. In one case, a biopsy was not repeated prior to tocilizumab and the patient was considered refractory based on persistent dysfunction and HLA DSA. Median (IQR) time from AMR diagnosis and tocilizumab treatment was 10.6 (8.3–17.6) months (Table 1), during which time the above described treatments were administered. These patients were offered treatment with tocilizumab as rescue therapy at the discretion of the physician as part of clinical care. There were no other inclusion or exclusion criteria.

Table 1:

Patient Characteristics

	Cohort (N=25)
Age (Years) at Transplant, Median (IQR)	8.6 (4.2–15.5)
Age (Years) at 1st Tocilizumab, Median (IQR)	17.9 (12.9–19.9)
Sex (Male)	14 (56.0%)
Ethnicity (Hispanic)	12 (48.0%)
Race
White	15 (60.0%)
Black	3 (12.0%)
Asian	1 (4.0%)
Other	6 (24.0%)
First Transplant	22 (88.0%)
Panel Reactive Antibody >20%***	2 (8%)
Deceased Donor	18 (72.0%)
Induction Immunosuppression
IL-2 Inhibitor	20 (80.0%)
ATG	3 (12.0%)
Alemtuzumab	2 (8.0%)
Maintenance Immunosuppression
FK, MMF, Prednisone	24 (96.0%)
FK, Sirolimus, Prednisone	1 (4.0%)
Immunosuppression Risk Prior to AMR Diagnosis
Medication Nonadherence Suspected/Reported	6 (25%)
Physician Directed Medication Reduction	6 (25%)
Pharmacy Error	1 (4%)
Unknown	12 (48%)
Time from transplant to AMR (Months), Median (IQR)	41.4 (24.3–67.7)
Time from AMR to 1st Tocilizumab (Months), Median (IQR)	10.6 (8.3–17.6)
Time from Transplant to Pre-Tocilizumab Biopsy (Months), Median (IQR)	58.2 (31.5–109.9)
Time from Pre-Tocilizumab Biopsy to Tocilizumab Initiation (Months), Median (IQR)	1.2 (0.6–2.8)
Time from Tocilizumab to Last Follow-up (Months), Median (IQR)	15.8 (8.4–35.7)
Number of Doses of Tocilizumab, Median (IQR)	12.0 (6.0–19.0)
Last Follow-up Description
Tocilizumab Ongoing	15 (60.0%)
At time of last dose of Tocilizumab (no adverse event)	5 (20.0%)
At time of last dose of Tocilizumab (due to adverse event)	3 (12%)
Tocilizumab Ongoing but on hold	1 (4.0%)
Graft Loss	1 (4.0%)
eGFR at 1st Tocilizumab, Median (IQR) (ml/min/1.73m2)	67.3 (55.0–78.0)
HLA DSA Positive at diagnosis of AMR *	23 (92.0%)
Pre-Tocilizumab pulse steroids	21 (84%)
Pre-Tocilizumab IVIG	25 (100%)
Pre-Tocilizumab Rituximab	24 (96.0%)
Pre-Tocilizumab Bortezomib	8 (32%)
Pre-Tocilizumab Plasmapheresis	12 (48%)
HLA DSA at 1st Tocilizumab
DQ	13 (52.0%)
DR	4 (16.0%)
A1	1 (4.0%)
None**	7 (28.0%)
AMR Type at Time of 1st Tocilizumab
Active AMR	16 (64.0%)
Chronic Active AMR	9 (36.0%)
C4d Positive at Time of 1st Tocilizumab***	13 (52%)
Concurrent ACR at Time of 1st Tocilizumab***
None	14 (56.0%)
borderline	5 (20.0%)
1A	3 (12.0%)
2A	2 (8.0%)
Concurrent IVIG during Tocilizumab therapy	11 (44%)

^*The 2 patients negative for HLA DSA had high AT1R-Ab (>17 units/mL)

^**Of these patients, 5 had high AT1R-Ab (>17 units/mL), 1 had AT1R-Ab in the “at risk” range (12 units/mL), and 1 was not tested for non-HLA

^***1 patient missing PRA data. 2 patients missing full biopsy scoring and 1 missing ACR evaluation

AMR: antibody mediated rejection, IL-2: interleukin 2, ATG: anti-thymocyte globulin, FK: tacrolimus, MMF: mycophenolate mofetil, eGFR: estimated glomerular filtration rate, HLA DSA: human leukocyte antigen donor specific antibody, ACR: acute cellular rejection

Before the start of tocilizumab, all patients were tested for exposure to tuberculosis with a MTB-Quantiferon-Gold Plus ELISA blood test. All patients were on steroid based immunosuppression with tacrolimus and an anti-metabolite at both the time of AMR diagnosis and time of tocilizumab initiation. Patients were monitored for adverse effects during tocilizumab infusion. The dose of tocilizumab was 8mg/kg with a maximum of 800mg over 60 −100 minutes¹⁷. Premedication consisted of acetaminophen 10mg/kg (maximum 650mg), diphenhydramine 1mg/kg (maximum 50mg), and methylprednisolone 0.5–1mg/kg (maximum 100mg), administered 30 minutes prior to tocilizumab infusion. Concurrent ACR was treated with methylprednisolone of 10mg/kg/day (maximum 500mg) IV for 3 days. After steroid pulse, steroid doses were tapered back down to 0.07–0.1 mg/kg/day. No patients received anti-thymocyte globulin for rejection treatment.

After initiation of therapy, patients continued their standard immunosuppression without change in steroid dose, target tacrolimus troughs (5–7 ng/mL for patients on mycophenolate and 2–4 for patients on mTOR inhibitors), or antimetabolite dosage. There was no change in adherence monitoring. Patients were monitored for viral infections including Cytomegalovirus (CMV), Epstein-Barr virus (EBV), BK virus, and JC virus monthly. Viral tests for CMV, EBV, and BK virus were considered positive if they were above the quantifiable range. JC virus was considered positive if detectable. IgG levels were assessed monthly and patients with total IgG <600 were repleted with IVIg. Eleven patients were treated with concurrent IVIg as treatment for AMR (Table 1). Liver function tests and complete blood counts were monitored. Dosing was adjusted or held for neutropenia (ANC<1500), elevated liver enzymes (> 3 times upper limit of normal), and viremia.

2.2. Outcomes Assessment

Clinical characteristics and treatment outcomes were collected by retrospective chart review. Clinical outcomes were collected at the following time points: at initiation of tocilizumab, 6 months and 12 months following initiation of tocilizumab (where applicable), and last follow up. The last follow up date was determined as the date of last administered dose of tocilizumab and/or last available follow up as of November 2020 for patients in whom treatment was ongoing at that time. Patients were observed for up to 6 years from start of tocilizumab. Estimated glomerular filtration rate (eGFR) was determined in all patients <18 years of age at start of tocilizumab using the modified Schwartz formula²⁰ and for patients >18 years of age with the CKD-Epi formula²¹. Urinary protein to creatinine ratios were not systematically obtained, so we were unable to accurately assess changes in proteinuria in our analysis. Allograft loss date was defined as dialysis start date.

HLA antibody testing was done using Luminex (One Lambda, CA) and quantified by mean fluorescent intensity (MFI). Closest available HLA DSA results to the study time points within 3 months were collected. Mean fluorescence intensity (MFI) quantification data were summarized as weak, moderate, and strong based on center specific HLA laboratory cutoffs to allow better comparability of MFI results between centers. MFI ranges for each category for UCLA/Cedars-Sinai respectively were: weak 1,000–3000/2,500–4,999, moderate 3,000–8000/5,000–9,999, strong >8000/10,000. Change in both immunodominant DSA (iDSA), and the relative intensity score (RIS) were evaluated. The RIS was calculated by adding scores for all positive HLA DSA based on MFI category (weak = 2 points, moderate = 5 points, and strong = 10 points). We used RIS in addition to iDSA in our analysis to incorporate a cumulative measure of HLA DSA burden, as it may be important in the pathogenesis of AMR²². HLA DSA assessments at the time of starting tocilizumab and at last available follow up test were used for the primary analysis. Angiotensin II type 1 receptor antibody (AT1R-Ab) when tested was performed using enzyme-linked immunosorbent-based assay (One Lambda, CA) with >17 units/mL considered strongly positive and >10 units/mL considered “at risk.”

Biopsy reports for all 25 patients meeting initial entry criteria, were generated by each center’s renal pathology team and graded according to Banff 2019 criteria for ACR and Banff 2017 criteria for AMR, the latter with the modification that chronic glomerulopathy (cg) scores were based solely on light microscopy as not all specimens were examined with electron microscopy^4,23. The most recent biopsy prior to tocilizumab therapy was used as the baseline biopsy for analysis. One patient did not strictly meet Banff 2017 AMR criteria at time of tocilizumab initiation due to lack of HLA DSA with no non-HLA testing performed. The patient was treated as refractory AMR based on persistence of significant microvascular inflammation on biopsy with history of previously positive HLA DSA. Indirect immunofluorescence was performed for C4d using monoclonal anti-C4d antibody as previously described²⁴. Nineteen of 25 patients (76%) had follow-up renal biopsies available for assessment. In the case of multiple biopsies, the last available biopsy score while on tocilizumab treatment was used to assess change over time. This allowed for a standardized assessment of change within the cohort given the variable follow up time frames.

2.3. Statistical analysis:

A mixed-effect linear regression model with a patient-level random effect was used to evaluate change in eGFR over time. To see if this result differed by ACR status, we also tested for an interaction between time and presence of ACR at the start of treatment. Change in iDSA, RIS, and biopsy scores from baseline to last available follow-up was compared to a mean change of 0 using a two tailed t-test. P-values less than 0.05 were considered statistically significant. All analyses were completed in Stata/SE 15.1.

3. Results

3.1. Patient Characteristics

Twenty-five pediatric renal transplant patients with refractory active or chronic active AMR were treated with tocilizumab. Demographic and clinical characteristics of the cohort are described in Table 1. The majority of patients were adolescents (median age 17.9 years) at the time of tocilizumab initiation. Four patients were under the age of 10. The majority of patients (88%) were first kidney transplants and were unsensitized. No patients had pre-formed HLA DSA to the donor. Patients treated using this protocol had AMR with a median (IQR) of time post-transplantation to AMR of 41.4 (24.3–67.7) months. There was no formal assessment of medication nonadherence. There were 6 patients who were nonadherent in the time leading up to AMR diagnosis based on physician assessment or patient report, 6 patients who had immunosuppression recently reduced because of viral infection and one patient who was given the incorrect concentration of tacrolimus by pharmacy. The number of patients receiving pulse steroids, IVIg, rituximab, bortezomib, and plasmapheresis for rejection treatment prior to tocilizumab are shown. Notably, all patients received IVIg, and all but one received rituximab. The median (range) of pre-tocilizumab doses of medications was 1(1–3) for rituximab and 4 (4–9) for bortezomib. For patients who received plasmaphereses pre-tocilizumab, the median (range) number of sessions was 12 (5–26). A median (IQR) of 12 (6.0–19.0) doses of tocilizumab were administered with a median (IQR) follow up period of 15.8 (8.4–35.7) months. The majority of patients had active AMR (64%) at the time of tocilizumab initiation with the remainder having chronic active AMR (36%). Almost all patients were HLA DSA positive at time of initial AMR diagnosis (23/25, 92%). The remaining 2 patients had strong (>17 units/mL) AT1R-Ab. Seven patients (28%) had resolution of HLA DSA with initial (pre-tocilizumab) AMR therapy. Concurrent ACR was present in 40% of patients at the time of tocilizumab initiation and treated with pulse methylprednisolone.

Tocilizumab was ongoing at the time of study completion in 15/25 patients (60%), so last follow-up was November 2020. In 5 patients, tocilizumab was discontinued during the study time frame for reasons unrelated to therapy tolerability. Last follow up for these patients was at time of last tocilizumab dose. Of those 5 patients, 3 were transferred to adult care, 1 was discontinued due to futility in the setting of ongoing medication nonadherence, and 1 was transferred to adult care and converted to eculizumab. Three patients discontinued tocilizumab due to adverse events (see side effects and adverse events below) and last follow up was assessed at time of last tocilizumab dose. One patient had last follow-up at time of allograft loss and another had tocilizumab on hold at time of last follow-up due to JC viremia (Table 1).

3.2. Renal Function

Renal function was assessed in our cohort over time. The mean (95% confidence interval) of eGFR change per year of follow up time from time of AMR diagnosis to time of tocilizumab initiation was −3.12 (−6.36, 0) ml/min/1.73m²/year. The mean (95% confidence interval) eGFR change per a year of follow up time from start of tocilizumab to last follow up was −1.43 (−5.57, 2.72) ml/min/1.73m²/year and −1.8 (−6.12, 2.52) ml/min/1.73m²/year excluding and including the patient with allograft loss respectively. The eGFR did not significantly worsen during this timeframe in our cohort, indicating stabilization of renal function following treatment with tocilizumab (Figure 1). We also completed a sub-analysis to examine any differential effect of time on eGFR based on the presence or absence of concurrent ACR at time of tocilizumab initiation. We did not find a differential effect of time by presence of concurrent ACR, suggesting that patients both with and without concurrent ACR had stable renal function over the follow up period (Figure 1). There was one allograft loss during the pre-specified follow up period. The patient started tocilizumab with an eGFR 37ml/min/1.73m². He received 2 doses of tocilizumab, which was then held for 4 months following development of low-grade (<500 copies/mL) BK viremia. He did not have rebound of his HLA DSA while off tocilizumab. He had interval admissions for fever (rule out line infection) and severe diarrhea with acute kidney injury. The family endorsed compliance with his medications during this interval and his drug levels suggested he was taking his medications. At the time of restarting tocilizumab his eGFR ranged from 29–34 ml/min/1.73m². However, when the patient came for his 4^th dose, his eGFR was 15ml/min/1.73m² and he required dialysis approximately 1 month later.

Figure 1: Change in eGFR over time after Treatment with Tocilizumab.

Patients with AMR and mixed AMR and ACR at the time of tocilizumab initiation are shown in panel a and b. Mean eGFR in the cohort over time is shown. eGFR was obtained at 6 months, 12 months, and at time of last follow up (median 15.8 months as described in Table 1). 10 patients had last follow up points at 20 months post-tocilizumab initiation or later. The impact of time on change in eGFR was assessed by mixed effects model and was not significant; this indicated that eGFR was stable in our cohort over time. This remained true on sub-analysis of patients with and without concurrent ACR.

AMR: antibody mediated rejection, eGFR: estimated glomerular filtration rate, ACR: acute cellular rejection

3.3. Antibody Characteristics

The impact of tocilizumab on strength of HLA DSA was assessed using both iDSA and RIS score. Of the 25 patients, 23 (92%) had HLA DSA at AMR diagnosis, and 18/25 (72%) still had HLA DSA after receiving other therapy for AMR before the start of tocilizumab therapy. The iDSA was to DQ or DR in 17/18 (94%) cases (Table 1). Of the 7 patients without HLA DSA at the time of tocilizumab initiation, 5/7 (71%) remained free of HLA DSA. Of the 2 patients who developed HLA DSA on tocilizumab, one had recurrence of a previously identified HLA DSA and the other developed a de novo HLA DSA. The patient who developed de novo HLA DSA received every other month tocilizumab therapy due to the presence of persistently positive coccidiomycosis titers prior the kidney transplant. Notably, 5 of these patients had high AT1R-Ab (>17 units/mL), 1 had AT1R-Ab in the “at risk” range (12 units/mL), and 1 was not tested for non-HLA at time of tocilizumab initiation.

There was either no change or a decrease in iDSA and RIS in 23/25 (92%) and 21/25 (84%) patients respectively (Figure 2). On subset analysis of the 18 patients with HLA DSA at the start of tocilizumab, there was a significant decrease in iDSA (p=0.049), and a trend towards a decrease in RIS (p=0.065) (Figure 2). We did not observe any impact on AT1R-Ab levels.

Figure 2: Change in HLA DSA after Treatment with Tocilizumab.

Change in HLA DSA was assessed at time of tocilizumab initiation and at last available follow up test while on tocilizumab therapy. The change in immunodominant DSA (iDSA) (a) and relative intensity score (RIS) (b) in the total cohort of 25 patients are shown. (a) iDSA antibody strength as assessed by MFI category (strong, moderate, or weak) is shown on the Y axis. iDSA strength before tocilizumab and at time of last follow-up is represented for each patient. The black line represents 11 patients who had strong iDSA both before and after tocilizumab therapy. The gray line represents 5 patients who were HLA DSA negative both before and after tocilizumab. The dashed lines each represent one patient and show the change in each of their iDSAs. (b) The relative intensity score was calculated by adding scores for all positive HLA DSA based on MFI (weak = 2 points, moderate = 5 points, and strong = 10 points). RIS before and at time of last follow up for each patient is represented. The gray line represents 5 patients who were HLA DSA negative both before and after tocilizumab. The dashed lines each represent one patient and show the changes in their RISs. There was no statistically significant change in strength of iDSA (p=0.520) or RIS (p=0.190) in the total cohort of 25 patients. Eighteen of the 25 patients in our cohort, after other treatments for AMR, remained positive for HLA DSA at the time of tocilizumab initiation. In this subset of patients, there was a statistically significant decrease in iDSA (p=0.049), but not RIS (p=0.065) with tocilizumab therapy.

3.4. Biopsy Findings

Changes in biopsy findings were also assessed. Follow up biopsies were available in 19/25 patients (76%). The median (IQR) time from start of tocilizumab to last available follow up biopsy was 396 days (396, 412.5). There was no statistically significant change in the presence or absence of ACR or AMR. However, there was a statistically significant reduction in peritubular capillaritis (ptc) (p=0.015) and C4d scoring (p=0.009) (Figure 3). There was no significant impact on scores for glomerulitis (g). We also assessed development of chronic findings. There was a statistically significant increase in transplant glomerulopathy (cg) score (p=0.018) over the follow up period, but no increase in interstitial fibrosis and tubular atrophy score (IFTA) (Figure 3). We did not observe any impact of initial ACR status on change in IFTA (p=0.84).

Figure 3: Change in Biopsy Findings After Treatment with Tocilizumab.

Change in biopsies was assessed by Banff scoring at time of starting tocilizumab and at last available follow up. Follow up biopsies were available in 19 of 25 patients. The median (IQR) time between biopsies was 396.0 days (213.5–861.5). There was a statistically significant reduction in C4d (a) and peritubular capillaritis (c) scores. There was no change in glomerulitis (b). For chronic changes, there was no change in IFTA (d), however, there was a statistically significant increase in transplant glomerulopthy (e). None, mild, moderate, and severe IFTA were assigned scores of 0,1,2, and 3 respectively for analysis.

cg, allograft glomerulopathy score, IFTA, interstitial fibrosis and tubular atrophy score; g, glomerulitis; ptc, peritubular capillaritis

3.5. Side Effects and Adverse Events

Tocilizumab was generally well tolerated. The frequency of adverse events was recorded (Table 2). Four patients had viremias in the quantifiable range: 1 EBV and 3 BK virus, all asymptomatic. There was no CMV viremia in the quantifiable range. The EBV viremia was low grade and resolved without intervention. For 2 patients with BK viremia, mycophenolate mofetil (MMF) was discontinued and leflunomide was prescribed until the resolution of viremia. One patient switched from MMF to leflunomide and then to everolimus due to persistent low grade BK viremia. In all 3 patients, tocilizumab was held until the resolution of BK viremia. No patients developed BK virus nephropathy or a BK viral load >10,000 copies/mL. One patient had tocilizumab held due to a norovirus infection and later for asymptomatic JC viremia. Two additional patients discontinued tocilizumab after the third dose due to fatigue. Blood count abnormalities were seen in 21 patients with 9 patients having anemia. Persistent thrombocytopenia in one patient necessitated changing the frequency of infusions to once every 6 weeks. Two patients with neutropenia had resolution with no intervention, while 2 had doses held and 1 required filgrastim. Transaminitis in 3 patients was transient. One patient had a neo-bladder rupture during tocilizumab treatment, which was successfully repaired surgically. This patient had a history of intestinal bladder augmentation and developed severe abdominal pain after her 12^th dose of tocilizumab. She was immediately admitted to the hospital where imaging revealed a bladder rupture. Tocilizumab treatment was discontinued at that time.

Table 2:

Side Effects/Adverse Events after Starting Tocilizumab

Event	N (%)	Comments	Months Post-Tocilizumab
Asymptomatic CMV viremia	0(0)
Asymptomatic EBV viremia	1(4%)	Persistent low-grade viremia starting prior to tocilizumab	preexisting
Asymptomatic BK viremia	3(12%)	Tocilizumab held until resolution of viremia. No cases of BK nephropathy. 2 patients: on Leflunomide 1 patient: converted from MMF to everolimus	3.2, 28.3, and 14.3
Mild transaminitis (AST,ALT< 100 U/L)	3(12)	Resolved without intervention	9.2, 0.5, and 6.4
Leucopenia (< 4×103/uL)	6(24)	Resolved without intervention	Mean (SD), 10.4 (8.0)
Neutropenia (<15 × 103/uL)	4(16)	2 patients: resolved without intervention 1 patient: intermittent filgrastim for ANC < 500 and doses held 1 patient: tocilizumab held for 2 weeks and then resumed	7.1, 15.3, 13.5, 11.8
Thrombocytopenia (< 100 × 103/uL)	2(8%)	1: resolved without intervention 1: Frequency of tocilizumab changed to every 6 weeks	0.3, 1.0
Anemia (< 35% Hematocrit)	9(36)	8 patients: no intervention 1 patient: MMF decreased by 25%	Mean (SD), 9.3 (11.6)
Other: Neobladder rupture	1(4)	Resolved with surgery Tocilizumab discontinued	20.5
Other Infectious Complication	1(4)	Patient with both norovirus and JC viremia Tocilizumab held	16.5, 39.2
Fatigue	2(8)	Tocilizumab discontinued	2.3, 2.3

CMV, Cytomegalovirus; EBV, Epstein-Barr Virus; AST, Aspartate Aminotransferase; ALT, Alanine Aminotransferase;

4. Discussion

This is the largest study to examine the use of tocilizumab as a rescue therapy for AMR in pediatric kidney transplant patients. We observed stabilization of renal function without significant changes in HLA DSA in our cohort. Interestingly, we also found reductions in both peritubular capillaritis and C4d staining and no progression of interstitial fibrosis and tubular atrophy.

Chronic AMR, and independently, the development of HLA DSA are associated with premature graft loss, steeper annual decline in eGFR, and reduced patient survival^25–28. Treatment options have focused on removal of circulating antibodies concurrent with reduction of their redevelopment through the use of plasmapheresis, anti CD20 (anti-B cell) monoclonal antibodies (ie, rituximab or ofatumumab), proteasome inhibitors (bortezomib), anti-C5 antibody (eculizumab) and IVIg¹⁹. Inhibition of IL6/IL6R has been shown in small studies to slow AMR progression because of its effects on B cell immunity (including blocking plasma cells), reduction in both total IgG and IgG1/3, and alteration of the balance between effector and regulatory T cells^29–31. In pediatric patients with AMR, high dose IVIg and Rituximab has been associated with a decline in detected HLA DSA, improved renal function, and histologic improvement (decreased C4d staining)^32,33. Stabilization of function in pediatric patients with refractory AMR using a protocol combining rituximab, bortezomib, plasmapheresis, and IVIg has also been described^18,34. However, this protocol has the disadvantage of requiring plasmapheresis, which can be challenging in patients with limited access options. Furthermore, bortezomib has subsequently been shown to be ineffective treatment for late AMR in a randomized controlled trial in adults³⁵. In pediatric patients with AMR that is refractory to IVIg and rituximab, our study describes the first published cohort treated with tocilizumab. Tocilizumab was well tolerated, stabilized graft function, and was associated with mild improvements in allograft histology, without requiring plasmapheresis.

Patients in our cohort had stable renal function while being treated with tocilizumab, with the exception of one allograft loss. This was true in patients both with and without concurrent ACR. Follow up times were variable, with some patients continuing to have stable renal function for over 36 months, and a mean (95% confidence interval) annual decline in eGFR of −1.8 (−6.12, 2.52) ml/min/1.73m²/year while on tocilizumab therapy. The development of de novo HLA DSA (dnDSA) alone has been associated with accelerated decline in eGFR over time. Weibe et. al have demonstrated in a cohort of 508 renal transplant recipients with an average annual decline in eGFR of −1.39 +/− 3.35 mL/min/1.73m²/year, that patients with dnDSA had a significantly higher rate of eGFR decline of −3.63 mL/min/1.73m²/year²⁸. The latter is similar to the annualized eGFR loss of −3.12 ml/min/1.73m²/year that we found in our cohort prior to tocilizumab treatment. Furthermore, active AMR has been associated with poor allograft survival, with 26% of patients experiencing allograft loss within 3 years³⁶. Patients with late chronic active AMR are at particularly high risk of allograft loss, with some studies showing a median allograft survival of only 1.9 years following diagnosis³⁷. In our pediatric cohort with refractory AMR treated with tocilizumab, stable renal function may reflect impact of treatment given the expected natural history of this condition. However, it is not possible to directly assess this given this was not a randomized controlled trial.

Previous studies using tocilizumab for desensitization showed rapid reductions of DSA. This group has also shown overall decreases in iDSA MFI values up to 24 months post initiation of tocilizumab therapy when used for the treatment of AMR¹⁷. More recently in 2020, Lavacca et al studied the effects of tocilizumab as a first line agent in a group of 15 patients with chronic AMR that were naïve to other therapies. The authors found that after tocilizumab use kidney function stabilized, mean DSA MFI decreased, and there was reduction in inflammation on biopsy³⁸. In our cohort of 25 patients, the majority of patients had either no change or decrease in iDSA (92%) and RIS (84%). On subset analysis of the 18 patients with HLA DSA at the start of tocilizumab (but after other AMR treatment), there was a significant decrease in iDSA (p=0.049), and borderline significant decrease in RIS (p=0.065) with tocilizumab treatment (Figure 2). Overall, however, we did not see a dramatic change in HLA DSA in our cohort. The mechanism of tocilizumab in preserving allograft function may go beyond decrease in antibodies. This may be explained by increases in T-regulatory cells or decreased IL-6 production, which have been shown in previous studies to be contributory³⁹,

Nineteen of our patients had follow up biopsy data a median time of approximately 1 year after starting tocilizumab therapy. We found statistically significant improvements in findings of active AMR including peritubular capillaritis and C4d score. This is consistent with the largest previous case series which showed improvement in microvascular inflammation (g+ptc score) and C4d scores on 1 year follow up biopsies in patients with refractory AMR¹⁷. In a smaller series of patients treated with tocilizumab as initial therapy for AMR, Lavacca et al. reported glomerulitis improved on 6 month follow up biopsies and there was no progression of transplant glomerulopathy or IFTA³⁸. In our series, we found an improvement in ptc, but not g scores. Further, on assessment of chronic findings, there was no change in IFTA score suggesting a lack of progression in terms of fibrosis. However, there was progression in cg score. Our data is internally consistent in that there was a larger potential impact of tocilizumab on disease progression in the tubulointerstitial compartment rather than in the glomerulus. Larger studies are needed to assess this pattern, as it may provide insight to drive further mechanistic studies and help identify patients that stand to benefit most from therapy.

Though tocilizumab was generally well tolerated, our experience highlights the need for careful monitoring during therapy. Several patients in our cohort had doses held or therapy discontinued (3/25, 12%) related to side effects (Table 2). Neutropenia is the most common side effect reported with tocilizumab use for autoimmune disease in adults with no increased risk of infections overall⁴⁰. These findings are similar to the adverse effects in our cohort, with the most common side effect of tocilizumab being bone marrow suppression. In a kidney transplant cohort comparing infection rate in patients receiving IVIg/Rituximab to patients receiving tocilizumab for the treatment of AMR, there was a lower risk of infections in the latter group⁴¹. Twenty percent of our patients had viral infections: 1 EBV viremia, and 3 with BK virus. BK viremia required holding tocilizumab and adjustments to immunosuppression. None of the patients had CMV or EBV disease or BK nephropathy.⁴⁰ The perforation of the neo-bladder may or may not be related to the use of tocilizumab; however, the drug has been associated with increased GI perforation in adult patients with rheumatoid arthritis⁴². We recommend prescreening patients for gastrointestinal issues and closely monitoring for these complications with tocilizumab use. This is also an exclusionary requirement for the ongoing clazakizumab (genetically engineered anti IL6 monoclonal antibody) trial for treatment of chronic AMR (NCT03744910)^43–45.

There are several limitations in this study. This was an observational study without a control arm and our cohort, like in all studies involving pediatric patients, was small. The study was conducted at 2 separate institutions where different renal pathologists scored and read the biopsy specimens. The measurements of HLA DSA MFI were also performed at different immunogenetics laboratories. Assessment using laboratory classifications of weak, moderate, and strong were used to mitigate the impact of laboratory variations in MFI. No molecular diagnostic criteria for AMR was included. No distinction was made between clinical or subclinical AMR. We also did not stratify based on chronic active AMR versus active AMR alone. Our analyses are based on retrospective collection of clinical data, and therefore the timing intervals for our outcomes assessments could not be completely standardized. Based on these limitations, the benefits of tocilizumab on pediatric kidney transplant AMR should be interpreted cautiously. A prospective, randomized clinical trial is needed to determine long term benefits, safety and efficacy of tocilizumab in pediatric kidney transplant patients with AMR.

Our case series adds to the current literature by describing the long-term safety and outcomes of tocilizumab in pediatric patients with refractory AMR. We found tocilizumab to be well tolerated by most patients in our cohort. Though we did not observe a significant decrease in HLA DSA, we found improvement in ptc and c4d scoring despite our small sample size. Renal function remained stable in our patients. This case series highlights the potential utility of tocilizumab for the treatment of AMR in pediatric kidney transplantation.

Abbreviations:

AMR

Antibody mediated rejection

IL-6

Interleukin-6

IQR

Interquartile range

DSA

Donor Specific Antibody

ACR

Acute Cellular Rection

HLA

Human Leukocyte Antigen

IVIg

Intravenous Immunoglobulin

MFI

Mean Fluorescent Intensity

CMV

Cytomegalovirus

EBV

Epstein-Barr Virus

ANC

Absolute neutrophil count

eGFR

Estimated Glomerular Filtration Rate

RIS

Relative Intensity Score

AT1R

Angiotensin II type 1 Receptor

Cg

Chronic glomerulopathy

iDSA

Immunodominant Donor Specific Antibody

ATG

Anti-thymocyte Globulin

MMF

Mycophenolate Mofetil

FK506

Tacrolimus

AST

Aspartate Aminotransferase

ALT

Alanine Aminotransferase

IFTA

Interstitial fibrosis and tubular atrophy

PTC

Peritubular Capillaritis

Data Availability Statement:

The data that support the findings of this study are available from the corresponding author upon reasonable request.