Infliximab for Grade III or IV Immune Checkpoint Inhibitor Nephritis Clinical and Translational Evidence

Abstract

Introduction

Acute interstitial nephritis (AIN) is a common presentation of immune checkpoint inhibitor (ICI) toxicity. Increased tumor necrosis factor (TNF)-α expression has been observed in ICI-induced AIN (ICI-AIN). In this study, we compared treatment with glucocorticoids plus infliximab to glucocorticoids only in stage III or IV ICI-AIN. We conducted retrospective NanoString analysis of ICI-exposed kidney samples to identify associated gene signatures and possible association with TNF-α expression.

Methods

We reviewed the records of patients who received ICI, developed acute kidney injury (AKI), and underwent kidney biopsy at MD Anderson Cancer Center from January 2022 to October 2024. Fifty-five patients had AIN and stage III or IV AKI. 27 received steroids only and 28 received steroids plus infliximab. Time to renal response (TTR), duration of renal response (DORR), and cancer progression–free survival (PFS) were estimated using the Kaplan-Meier method. The log-rank test assessed survival differences between groups. Cox proportional hazards analyses were conducted on TTR and DORR, with treatment as a time-dependent covariate for TTR. Using NanoString technology, inflammatory gene signatures were evaluated in 86 kidney samples, of which 48 were ICI-exposed.

Results

No baseline differences were observed between groups. TTR was significantly improved in patients exposed to infliximab within 30 days of starting steroids compared with steroid-only. DORR was longer in infliximab group in univariate but not multivariate analysis, and no PFS difference was observed. In NanoString analysis, immune expression was higher in drug- and ICI-AIN compared with controls. Elevated TNF-α expression in drug- and ICI-AIN correlated with renal response.

Conclusion

Infliximab use may offer a therapeutic advantage over steroids alone in stage III or IV ICI-AIN.

Graphical abstract

ICI-associated AKI is a well-described immune-related adverse event (irAE) encountered in 3% to 5% of patients treated with ICIs.^1,2With the rising indications for ICI use, the incidence of ICI-associated AKI continues to increase. Several renal histopathological findings have been described in association with ICI, with ICI-AIN being the most reported (75%–80%).³ In the absence of prospective clinical data, treatment of ICI-AIN can be challenging. To date, steroid therapy and temporary cessation of the offending ICI agent remain the standard of care for ICI-AIN.^4,5

Although steroid therapy is effective in most ICI-AIN cases, < 50% of patients achieve complete recovery of renal function.^6,7 Prolonged kidney injury plays a critical role in the development of tissue fibrosis and significantly reduces the likelihood of recovery.⁸ Furthermore, relapses, along with steroid-dependent and steroid-refractory ICI-AIN, pose considerable challenges and often require prolonged corticosteroid treatment.

Given the potential adverse effects of long-term corticosteroid use, including its impact on cancer treatment and sustained immunosuppression, along with the risk that prolonged kidney injury may hinder renal recovery, alternative therapeutic strategies are needed.

Elevated serum TNF-α levels have been observed in patients with ICI-AIN,9, 10 and urinary TNF-α levels are significantly higher in individuals with AIN of any etiology.11, 12 Given these findings, targeting TNF-α for ICI-AIN treatment has been explored. Infliximab, an anti–TNF-α monoclonal antibody, has shown promise as a treatment for relapsed ICI-AIN or as salvage therapy in patients with partial or no renal recovery following initial steroid treatment.^11,13 However, the use of infliximab as a first-line therapy for ICI-AIN has not been previously described.

In this retrospective study, we evaluated the outcomes of patients with biopsy-proven AIN and stage III or IV AKI secondary to ICI-AIN treated with infliximab plus steroids as first-line therapy. We compared their outcomes to those of patients treated with steroids alone. We focused on the time to response to therapy, associated AKI relapse rate, PFS, and overall survival (OS) from the time of ICI-AIN treatment initiation to death or the last time patient was known to be alive. In addition, we retrospectively evaluated available kidney samples using NanoString technology to investigate the associated immune environment in patients exposed to ICI compared with controls.

Methods

We retrospectively reviewed the medical records of all patients who received ICI treatment, developed AKI, and had a diagnostic kidney biopsy at The University of Texas MD Anderson Cancer Center from January 2022 to October 2024. This retrospective study was approved by MD Anderson’s Institutional Review Board in accordance with the Declaration of Helsinki. We identified 55 patients with biopsy-proven AIN who had stage III or IV AKI from January 2022 to October 2024. Of these, 27 were treated with steroids only (equivalent prednisone dose ≥ 0.5 mg/kg) and 28 were treated with infliximab (5 mg/kg) in addition to steroids.

Clinical data and laboratory results were reviewed for each patient. We extracted the following baseline information for each patient: age, gender (male/female), cancer diagnosis, comorbidities, name and class of ICI, potentially nephrotoxic medications (proton pump inhibitors, nonsteroidal antiinflammatory drugs), urine sediment at the time of AKI, presence of concomitant nonrenal irAEs, and serum creatinine at baseline.

Outcome variables were the following: peak creatinine during AKI, need for renal replacement therapy, renal response at 3 months from treatment initiation (infliximab and/or steroids), PFS was assessed based on imaging as the best cancer response at time of renal response, and best cancer response at last follow-up. Overall response was defined as having stable disease, remission, or any improved tumor response; and “no response” was defined as any progression of disease. AIN diagnosis was made by a board-certified renal pathologist based on a renal biopsy pathological diagnosis.

In addition to TTR, we evaluated OS and PFS from the time from steroid initiation to death or the last time the patient was known to be alive.

Patients in our study were diagnosed with stage III or IV AKI based on common terminology criteria for adverse events V5.0 as follows: grade III AKI, > 3× baseline creatinine; grade IV, > 6× baseline creatinine or dialysis needed. Complete renal response was defined as post-AKI improvement in creatinine to < 0.35 mg/dl above baseline. Partial renal response was defined as creatinine > 0.35 mg/dl but less than the peak creatinine value at the time of AKI diagnosis.

Statistical Analysis

Patients were divided into 2 treatment groups as follows: the steroid-infliximab group included patients who received infliximab in addition to steroids, and the second group received only steroids. We evaluated the TTR and DORR in the entire cohort from the time of steroid initiation regardless of when infliximab was given during the study period, and another analysis was performed including only patients who received infliximab within 30 days of steroid initiation.

Descriptive statistics (frequency distribution, mean [SD], and median [range]) were used to summarize patients’ characteristics. Chi-square or Fisher exact tests for categorical variables; and t test, analysis of variance, or their nonparametric counterparts (Wilcoxon rank-sum or Kruskal-Wallis) for continuous variables were used to compare patient’s characteristics between groups.¹⁴ The distributions of TTR, DORR,OS, and PFS by renal response evaluation were estimated using the Kaplan-Meier method.¹⁵ TTR was the time from steroid initiation to response determined 3 months after the initial response evaluation. DORR was the time from renal response evaluated 3 months after the initial response evaluation to renal relapse or last renal function follow-up. Thus, DORR was only calculated for the patients who achieved a renal response. PFS was defined as the interval from steroid initiation to progression or death, whichever occurred first. For events that had not yet occurred, times were censored at the time of renal response evaluation (first censoring time) or last follow-up (second censoring time) at which the patient was known to be progression-free. OS was defined as the time from steroid initiation to death or the last time the patient was known to be alive. The log-rank test¹⁶ was performed to test the difference in survival between groups. Regression analyses of survival data based on the Cox proportional hazard model were conducted on TTR and DORR.¹⁷ In the TTR analysis, treatment was considered a time-dependent covariate. SAS version 9.4 (SAS Institute Inc., Cary, NC) and S-Plus version 8.2 (TIBCO Software Inc., Palo Alto, CA) were used for all analyses.

NanoString Analysis

We analyzed 86 kidney samples, of which 48 were treated with ICI. The 38 control samples included 11 with normal biopsy findings (no pathology, diabetes, and hypertension changes), 13 with non-ICI–related acute tubular necrosis (ATN), and 14 with drug-induced AIN. The ICI-exposed samples included 7 with no AIN, 8 with ATN, and 33 with AIN with or without ATN, all of which were a part of the cohort studied for therapeutic outcome (Figure 1). Samples were processed as previously described.¹⁸ RNA was isolated from formalin-fixed paraffin-embedded kidney sections by dewaxing using deparaffinization solution (Qiagen, Valencia, CA), and total RNA was extracted using the RecoverALL Total Nucleic Acid Isolation kit (Ambion, Austin, TX) according to the manufacturer’s instructions. The RNA purity and quantity were assessed on the ND-Nanodrop1000 spectrometer (Thermo Scientific, Wilmington, MA). For the NanoString assay, 100 ng of RNA was used to detect immune gene expression using nCounter PanCancer Immune Profiling panel along with custom CodeSet. Counts of the reporter probes were tabulated for each sample by the nCounter Digital Analyzer and raw data output was imported into nSolver (http://www.nanostring.com/products/nSolver). nSolver (advanced analysis 2.0) data analysis package was used for normalization, cell type analysis, and differential gene expression. Hierarchical clustering heatmap analysis and GSEA were performed using Qlucore Omics Explorer version 3.5 software (Qlucore, NY). Data were plotted using GraphPad Prism 10 (GraphPad Software v-10.4.2) and 2-tailed Mann-Whitney test was performed to compare 2 groups, P-value < 0.05 was considered significant. P-values were adjusted for multiple comparisons using the Benjamini-Hochberg Procedure. Welch’s t test (determined by nSolver advanced analysis 2.0 software) was used to determine P-values for volcano plot (differential expression) and P-value (FDR-adjusted < 0.2) were used for GSEA analysis.

Figure 1.

Nano string analysis cohort. AIN, acute interstitial nephritis; ATN, acute tubular necrosis; ICI, immune checkpoint inhibitor; QC, quality control.

Results

Patient Demographics and Clinical Characteristics

Baseline Characteristics

We identified a total of 55 patients who were diagnosed with cancer, treated with ICI, and subsequently developed grade III or IV ICI-AIN. Of these, 43% had grade IV AKI and 56% had grade III AKI. Patients’ clinical characteristics, including comorbidities, cancer type, and cancer-related history, are provided in (Table 1, Supplementary Table S1). Of the entire cohort, 61.82% were males and 38.18% were females, with a median age of 68 (Supplementary Table S1).

Table 1.

Patient’s demographic and clinical characteristics stratified by treatment group

Demographic and clinical characteristics		Patients, n (%)			Patients within 30 d of start of ICI-AIN treatment, n.(%)
Covariate	Levels	Steroid alone	Steroid + infliximab	P	Steroid alone	Steroid + infliximab	P
Gender	F	12 (57.1)	9 (42.9)	0.3479	15 (71.4)	6 (28.6)	0.4640
	M	15 (44.1)	19 (55.9)	.	21 (61.8)	13 (38.2)	.
HTN	No	8 (66.7)	4 (33.3)	0.2047	8 (66.7)	4 (33.3)	1.000
	Yes	19 (44.2)	24 (55.8)	.	28 (65.1)	15 (34.9)	.
Hyperlipidemia	No	16 (55.2)	13 (44.8)	0.3407	21 (72.4)	8 (27.6)	0.2517
	Yes	11 (42.3)	15 (57.7)	.	15 (57.7)	11 (42.3)	.
DM	No	22 (57.9)	16 (42.1)	0.1051	28 (73.7)	10 (26.3)	0.0803
	Yes	5 (31.3)	12 (68.8)	.	8 (50)	9 (50)	.
CAD	No	26 (57.8)	19 (42.2)	0.0116	32 (71.1)	13 (28.9)	0.0774
	Yes	1 (10)	9 (90)	.	4 (40)	6 (60)	.
Comorbidities	No	7 (77.8)	2 (22.2)	0.0776	7 (77.8)	2 (22.2)	0.4734
	Yes	20 (43.5)	26 (56.5)	.	29 (63)	17 (37)	.
Breast	No	27 (50)	27 (50)	1	36 (66.7)	18 (33.3)	0.3455
	Yes	0 (0)	1 (100)	.	0 (0)	1 (100)	.
GI	No	24 (49)	25 (51)	1	31 (63.3)	18 (36.7)	0.6525
	Yes	3 (50)	3 (50)	.	5 (83.3)	1 (16.7)	.
Lung	No	22 (47.8)	24 (52.2)	0.7287	30 (65.2)	16 (34.8)	1.000
	Yes	5 (55.6)	4 (44.4)	.	6 (66.7)	3 (33.3)	.
Thyroid	No	27 (51.9)	25 (48.1)	0.2364	34 (65.4)	18 (34.6)	1.000
	Yes	0 (0)	3 (100)	.	2 (66.7)	1 (33.3)	.
Hematologic	No	22 (44)	28 (56)	0.0232	31 (62)	19 (38)	0.1518
	Yes	5 (100)	0 (0)	.	5 (100)	0 (0)	.
Dermatologic	No	22 (53.7)	19 (46.3)	0.355	28 (68.3)	13 (31.7)	0.4488
	Yes	5 (35.7)	9 (64.3)	.	8 (57.1)	6 (42.9)	.
Genitourinary	No	22 (47.8)	24 (52.2)	0.7287	30 (65.2)	16 (34.8)	1.000
	Yes	5 (55.6)	4 (44.4)	.	6 (66.7)	3 (33.3)	.
Others	No	23 (47.9)	25 (52.1)	0.7049	32 (66.7)	16 (33.3)	0.6823
	Yes	4 (57.1)	3 (42.9)	.	4 (57.1)	3 (42.9)	.
Concurrent nephrotoxic chemo	No	15 (53.6)	13 (46.4)	0.4985	19 (67.9)	9 (32.1)	0.7028
	Yes	12 (44.4)	15 (55.6)	.	17 (63)	10 (37)	.
Concurrent ICI	No	18 (54.5)	15 (45.5)	0.3217	23 (69.7)	10 (30.3)	0.4177
	Yes	9 (40.9)	13 (59.1)	.	13 (59.1)	9 (40.9)	.
CTCAE grade	III	15 (48.4)	16 (51.6)	0.9055	21 (67.7)	10 (32.3)	0.6852
	IV	12 (50)	12 (50)	.	15 (62.5)	9 (37.5)	.
Dialysis	No	23 (56.1)	18 (43.9)	0.1212	30 (73.2)	11 (26.8)	0.0395
	Yes	4 (28.6)	10 (71.4)	.	6 (42.9)	8 (57.1)	.
UA findings, RBC, WBC	No	4 (44.4)	5 (55.6)	1	4 (44.4)	5 (55.6)	0.2489
	Yes	23 (50)	23 (50)	.	32 (69.6)	14 (30.4)	.
Steroids before biopsy	No	17 (70.8)	7 (29.2)	0.0045	20 (83.3)	4 (16.7)	0.0217
	Yes	10 (32.3)	21 (67.7)	.	16 (51.6)	15 (48.4)	.
Solumedrol	No	24 (55.8)	19 (44.2)	0.1011	30 (69.8)	13 (30.2)	0.2029
	Yes	3 (25)	9 (75)	.	6 (50)	6 (50)	.

AIN, acute interstitial nephritis CAD, coronary artery disease; CTCAE, common terminology criteria for adverse events; DM, diabetes mellitus; F, female; GI, gastrointestinal; HTN, hypertension; ICI, immune checkpoint inhibitor; M, male; RBC, red blood cells; UA, urinalysis; WBC, white blood cells.

Hypertension was the most reported comorbidity (78%), followed by hyperlipidemia (47%), and type 2 diabetes mellitus (29%). Half the patients (52%) had stage II chronic kidney disease (CKD) at baseline (Supplementary Table S1). The most common underlying malignancy type was dermatologic (25%), followed by genitourinary (16%) and lung (16%) (Supplementary Table S1).

Sixty percent of patients received single-agent ICI, whereas 40% received concurrent treatment with 2 ICIs, with a median of 4 cycles before ICI-AIN. Half of the patients (49%) had been exposed to nephrotoxic chemotherapies with ICI. Close to half the patients (47%) had other nonrenal irAE(s) reported, with gastrointestinal irAEs being the most common (18%). White and red blood cells were noted in the urine analysis of 83% of patients at the time of AKI. More than half of the patients (56%) were exposed to steroids before the biopsy, and 54% were receiving proton-pump inhibitors at the time of the AKI. Median baseline, peak, and final creatinine values were 1.0, 4.20, and 1.51 mg/dl, respectively.

Treatment-Related Variables

All patients were treated with steroids, whereas 28 patients (50.91%) received the steroid-infliximab combination. The number of infliximab doses ranged from 1 to 6, with a single dose being the most common, accounting for 34%. The median time from AKI diagnosis to steroid initiation was 5 days, and the median time to infliximab initiation was 28 days. Fourteen patients, 10 patients (35.7%) in the infliximab group and 4 patients (14.8%) in steroids-only group, required hemodialysis initiation. Fifty percent came off dialysis in both groups, 5 of 10 in the infliximab group, and 2 of 4 in the steroids-only group. Most patients (53%) experienced complete renal responses after treatment, 42% experienced partial responses, and 5% had no response. After the initial AKI, 16.36% of the patients were rechallenged with ICI. Notably, there were more patients in the steroid-infliximab group that were exposed to steroids before biopsy (P = 0.005), had an older median age (P = 0.003), and were more likely to be on dialysis at the time of AKI (57%; P = 0.0395) (Supplementary Table S1).

Time to Response

Patients treated with steroid plus infliximab had a significantly shorter time to response than patients treated with steroids only, with 33.3% versus 11.4% achieving renal response at 0.5 months and 61.1% versus 31.9% at 1 month, respectively (P = 0.0011, Figure 2). Patients treated with steroid and infliximab were more likely to respond to treatment than those treated with steroid alone (hazard ratio [95% confidence interval]: 2.448 [1.292–4.638], P = 0.0061). This was validated independently in a multivariate analysis controlled for CKD stage and dialysis at baseline (hazard ratio: 2.44 [1.292–4.638], Table 2) (Supplementary Figure S1).

Figure 2.

Renal response rate over time by treatment group.

Table 2.

Association between treatment or potential confounders and time to response in Cox proportional hazard models with treatment analyzed as a time-dependent covariate

Parameter	Description	Univariate				Multivariate
		Hazard ratio	95% confidence limits		P-value for overall effect	Hazard ratio	95% confidence limits	P-value for overall effect
Treatment	Steroid + infliximab vs. steroid alone	2.710	1.490	4.930	0.0011	2.448	1.292 4.638	0.0061
CKD grade, baseline	III or IV vs. I or II	1.256	0.642	2.458	0.5057	1.006	0.505 2.003	0.9874
Dialysis	Yes vs. no	2.019	1.057	3.858	0.0334	1.523	0.752 3.083	0.2427
PPI/NSAIDs/nephrotoxins	Yes vs. no	1.245	0.705	2.198	0.4507

CKD, chronic kidney disease; NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor.

Renal Relapse–Free Survival

In our evaluation of renal response over time, overall responders had a median follow-up time of 8.5 months, with a 6-month renal relapse–free rate (95% confidence interval) of 0.77 (0.55–0.89) and 0.53 (0.21–0.78) for the steroid-only and steroid-infliximab groups, respectively (P = 0.038), Supplementary Figure S2). In a multivariate analysis controlled for proton pump inhibitor or nonsteroidal antiinflammatory drug exposure, patients in the steroid-infliximab group were more likely to relapse compared with the steroid-only group. However, this was not significant in a multivariate analysis (P = 0.0968, hazard ratio: 2.751 [0.833–9.083], Table 3).

Table 3.

Association between treatment or PPI and duration of renal response in Cox proportional hazard models

Parameter	Description	Univariate				Multivariate
		Hazard ratio	95% confidence limits		P-Value for overall effect	Hazard ratio	95% confidence limits	P-Value for overall effect
Infliximab treatment (within 30 d)	Steroid + infliximab vs Steroid alone	2.923	1.012	8.443	0.0475	2.751	0.833 9.083	0.0968
PPI/NSAIDs/nephrotoxins	Yes vs. No	1.135	0.368	3.500	0.8254	0.805	0.237 2.736	0.7282

NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor.

PFS and OS

Our evaluation of PFS at the time of renal response showed no significant difference between the steroid-infliximab and steroid-only groups (P = 0.225, Supplementary Figure S3). In our evaluation of OS by renal response, the median follow-up was 2 months, and no patients died before renal response.

Recovery of CKD From Baseline to Last Follow-up

At baseline, 53% of patients had stage 2 CKD, based on glomerular filtration rate. At 3-months follow-up, glomerular filtration rate improvement was observed in only 25% of patients, with 73% showing worsening glomerular filtration rate and CKD stage. Among patients receiving infliximab within 30 days of steroid initiation, we found no difference in the percentage of patients with worse grade of CKD at 3 months after initial response evaluation than baseline between the steroid-only and steroid-infliximab groups (72.2% vs. 77.8%, P = 0.752, Supplementary Table S2).

Immune Profiling in ICI- and Drug-Induced AIN

Using NanoString transcriptomic profiling, we evaluated the immune landscape of kidney biopsies from patients with AIN, induced by ICI or other drugs, compared with controls and ATN cases. CD45⁺ immune cell infiltration was significantly elevated in both ICI-induced and drug-induced AIN compared with controls (P = 0.0003) and ATN (P = 0.005 in the drug-treated cohort; P = 0.0007 in the ICI-treated cohort), highlighting a robust immune response in AIN (Figure 3). However, no significant differences were observed between the 2 AIN subgroups (P = 0.216), suggesting a shared immune activation pattern regardless of etiology.

Figure 3.

Immune expression profile comparing ICI treated and control cohort. AIN, acute interstitial nephritis; ATN, acute tubular necrosis; ICI, immune checkpoint inhibitor; IL-6, interleukin-6; TLS, tertiary lymphoid structures; TNF, tumor necrosis factor.

Further classification revealed that CD8⁺ T cells, cytotoxic lymphocytes, macrophages, and B cells were significantly enriched in both ICI- and drug-induced AIN compared with controls and ATN (Figure 3). Notably, expression of gene signature associated with tertiary lymphoid structures was elevated in AIN samples compared with controls (P = 0.0336) and ATN (P = 0.0044) within the ICI-treated cohort, although no difference was observed between the AIN subgroups (P = 0.827), further suggesting that AIN exhibits a conserved immune architecture regardless of whether it is drug- or ICI-induced. Considering that bulk CD4 gene expression is not a reliable surrogate for CD4⁺ T cells, because myeloid subsets (particularly monocytes, macrophages, and dentritic cells) also express CD4 transcripts, we have excluded bulk CD4 expression from our analyses.

Both TNF-α and interleukin (IL)-6 were significantly upregulated in ICI- and drug-induced AIN samples compared with controls and ATN samples, indicating increased proinflammatory signaling (Figure 3). Consistent with this immune activation, PD-L1 expression was significantly increased in both forms of AIN compared with controls and ATN across cohorts (Supplementary Figure S4).

Among patients who experienced renal response, TNF-α expression trended higher than nonresponders (P = 0.070), whereas IL-6 expression was not significantly different (P = 0.444, Figure 4a). Differential gene expression analysis using volcano plots revealed upregulated inflammatory pathway genes, including TNF-α, in patients who experienced renal response compared to nonresponders (Figure 4b), indicating that therapeutic modulation of the heightened inflammatory milieu may be associated with renal recovery. Lastly, TNF-α expression did not differ significantly between patients with lower-grade (I or II) versus higher-grade (III or IV) kidney toxicity (P = 0.564), suggesting that TNF-α levels may reflect an immune activation state rather than toxicity severity (Supplementary Figure S5).

Figure 4.

(a) TNF-α expression depending on renal response (left) and IL-6 expression depending on renal response (right). (b) Volcano plot expression of inflammatory versus inhibitory genes in patients who were treated with immune checkpoint inhibitor and had renal response versus no renal response. IL-6, interleukin-6; NR, no response; R, response. TNF, tumor necrosis factor.

These findings reinforce the presence of a shared, robust immune response in both ICI- and drug-induced AIN, characterized by increased immune cell infiltration and cytokine expression. The comparable expression profiles between AIN etiologies suggest a conserved inflammatory program, with potential therapeutic implications. The association between TNF-α expression and renal response provides mechanistic support for the clinical utility of TNF-α blockade, such as infliximab, in managing ICI-related nephritis.

Discussion

This retrospective study introduces TNF-α inhibitors as a possible alternative to prolonged steroids therapy to treat ICI-induced AIN on the tissue level, where higher TNF-α expression was associated with renal response. Elevated serum and urinary levels of TNF-α have been noted in patients with ICI ATN.^12,19 This was supported by a study by Farooqui et al.¹⁰ that included 24 patients presenting with AKI during ICI therapy. The participants with ICI-related AKI had higher urinary TNF-α levels than those with non-ICI–related AKI. Given TNF-α’s role in inflammation, TNF inhibitors are commonly used to treat inflammatory diseases such as rheumatoid and psoriatic arthritis and inflammatory bowel diseases.20, 21, 22 Infliximab has been used as a second-line treatment for different irAEs, including pneumonitis, colitis, and myocarditis.²³ Evidence of infliximab for treating AIN is limited, with a few case reports of successfully treated sarcoid and granulomatous interstitial nephritis.^24,25 The use of infliximab in the setting of ICI-AIN was first reported in a cohort of 10 patients who experienced relapse within 4 weeks of steroid cessation or who had only partial or no renal recovery with corticosteroid therapy. Four of these patients achieved complete renal recovery, 4 had partial recovery, and 2 showed no improvement.¹¹

Another recent study described 3 cases of steroid-dependent ICI-AIN in which infliximab was effective as a steroid-sparing agent. In these cases, infliximab dosing was guided by both serum (C-reactive protein, TNF-α, sIL-2R) and urinary biomarkers (TNF-α/Cr, CXCL9/Cr, RBP/Cr). Notably, a decline in these biomarkers coincided with improvement in kidney function, suggesting a possible role for biomarker-driven infliximab therapy.¹³

In our study, patients receiving a first-line combination of steroids and infliximab showed a higher likelihood of responding to therapy than those treated with steroids alone. This finding remained statistically significant in a multivariate analysis, even after adjusting for CKD stage and the need for dialysis. However, when we evaluated durability over time, patients who were treated with infliximab and steroids were more likely to develop renal relapse than patients treated with steroids only. At 6 months follow-up, 77% of patients in the steroid-only group and 53% of the steroid-infliximab group were relapse-free. This finding might be related to possible selection bias, because we tend to give infliximab as a first-line treatment for patients with more severe features.

Time to Intervention or Response

Our findings support the potential benefit of early intervention with infliximab in AIN to maximize the likelihood of renal recovery. Previous studies have linked renal recovery to both the duration of kidney injury and the degree of interstitial fibrosis. A study by Praga et al. showed that irreversible fibrotic transformation mediated by profibrotic cytokines and growth factors occurs within just 7 days of active inflammation.⁸ Additional reports demonstrated higher renal recovery rates and improved OS with early treatment of AIN.^4,26 Similarly, in a previous study by our group, infliximab therapy within 2 months of ICI-AIN diagnosis was associated with higher likelihood of complete renal recovery than later treatment.²⁷ These findings underscore the potential benefits of early infliximab use, as seen in our cohort, where a shorter time to response was associated with improved outcomes.

Safety of Infliximab in the Oncologic Setting

However, the safety of TNF inhibition in the setting of active cancer therapy remains a concern. TNF-α’s role as protumor and antitumor mediator has been explored in several studies, with overall evidence supporting protumorigenic effect. Although it is involved in cytotoxic T-cell–mediated tumor cell death and the suppression of regulatory T and B cells, TNF-α has been shown to promote tumor growth via angiogenesis and inflammatory signaling. In murine models, TNF blockade enhanced antitumor immunity and overcame resistance to anti–PD-1 therapy.²⁸ Similarly, development of TNF knockout mice proved that lack of TNF was protective against skin tumors.^29,30 Furthermore, TNF inhibition had a protective effect against chemically induced lung and colon tumors in mice.^31,32

Nevertheless, clinical results have been inconsistent. Some studies suggest a negative impact on tumor response,^33,34 whereas others report favorable outcomes.^11,23,35, 36, 37, 38, 39 In the setting of irAEs, evidence is conflicting regarding the safety of infliximab use. Several ICI-induced enterocolitis studies have shown no survival difference between infliximab-treated and non-infliximab–treated groups,^36,40,41 whereas 1 study reported a significantly lower survival rate in patients treated with steroid plus TNF blockade for steroid-refractory irAEs compared with patients treated with steroid alone.³³ These results might have been confounded by factors such as steroid dose and duration.

In a previous human study, 7 of 10 patients treated with infliximab for ICI-AIN had stable disease without evidence of cancer progression, whereas only 3 experienced disease progression; PFS ranged from 8 to 20 months. In the same study, 4 patients who achieved complete kidney recovery demonstrated complete tumor regression and no progression at last follow-up.¹¹ However, in our study, no oncologic benefit (PFS, OS) was demonstrated. Another concern would be de novo formation of autoantibodies such as antinuclear antibodies, anti–double-stranded DNA antibodies, and anticardiolipin antibodies which has been reported in patients with rheumatoid arthritis after long-term use of infliximab.^42,43 In our cohort, most patients had brief exposure to infliximab, a single dose, with only 1 reported case of autoantibody development. Large-scale clinical studies are needed to clarify the risks and benefits of infliximab use in patients with ICI-AIN.

The novel finding in our study is the translational work (NanoString analysis) showing clear evidence of higher TNF-α expression in ICI- or drug-induced AIN. We have demonstrated significantly higher expression of TNF-α (ICI-AIN, P = 0.0005) and IL-6 (ICI-AIN, P = 0.0006) compared with control samples (no ICI exposure). However, we observed elevated expression of TNF-α but not IL-6 in ICI-exposed patients who experienced renal response. Because previous findings showed increased CD4⁺ T cell infiltration in both ICI-AIN¹⁰ and drug-induced AIN,⁴⁴ we have excluded bulk CD4 expression from our analyses. In conclusion, the gene expression among patients with ICI-AIN and renal response was proinflammatory, and fewer inhibitory pathways were expressed than ICI-exposed patients without a renal response.

Limitations

Our study has several limitations. First, it is a retrospective, single-center study, with a small sample size, which limits the generalizability of our findings. We could not standardize the timing of treatment initiation across the groups, which may introduce variability. In addition, in the absence of current guidelines, patients’ selection to receive infliximab was based on treating clinician evaluation, which contributes to selection bias. Moreover, infliximab was given to patients with more severe kidney injury, which might have led to confounding by indication bias. Data on cancer stages and the specifics of different cancer treatments were not fully available for retrospective analysis, potentially affecting the completeness of our findings, especially in the PFS and OS analyses. Finally, the NanoString analysis would be more complete with matching tumor tissue to investigate the tumor microenvironment and tumor response. Such analyses are underway.

Conclusion

Our findings indicate that first-line infliximab use may offer a therapeutic advantage over corticosteroids alone in patients with stage III or IV ICI-AIN, with a shorter TTR. This is an observational study with hypothesis-generating findings. Future prospective studies are needed to validate these findings and investigate infliximab’s long-term efficacy and oncologic safety.

Disclosure

PS is a member of the Scientific Advisory Board for each of the following companies: Achelois, Adaptive Biotechnologies, Akoya Biosciences, Apricity, Asher Bio, BioAtla LLC, BioNTech, Candel Therapeutics, Catalio, C-Reveal Therapeutics, Dragonfly Therapeutics, Earli Inc, Enable Medicine, Glympse, Henlius/Hengenix, Hummingbird, ImaginAb, InterVenn Biosciences, JSL Health, LAVA Therapeutics, Lytix Biopharma, Matrisome, Neuvogen, NTx, Oncolytics, Osteologic, PBM Capital, Phenomic AI, Polaris Pharma, Soley Therpeutics, Sporos, Spotlight, Time Bioventures, Trained Therapeutix Discovery, Two Bear Capital, Vironexis, and Xilis Inc. AD reported consultation fees for Amgen. All the other authors declared no competing interests.

Supplementary Material

Supplementary File (PDF)

Figure S1. Swimmer plot illustrating AKI onset, steroid and infliximab use, clinical response, and relapse events. Figure S2. Kaplan-Meier curve of renal relapse rate by treatment group. Figure S3. Kaplan-Meier curve of progression-free survival by treatment group at the time of renal response. Figure S4. PD-L1 expression comparing ICI-treated to non–ICI-treated patients. Figure S5. Comparing TNF-α expression depending on grade of kidney toxicity. Table S1. Patient’s demographic and clinical characteristics. Table S2. Chronic kidney disease recovery by treatment arms. STROBE checklist.