Amphiregulin as a biomarker for acute GvHD

During the 63rd American Society of Hematology (ASH) Annual Meeting & Exposition, Shernan Holtan presented a talk on the use of amphiregulin (AREG) as a monitoring biomarker during acute graft-versus-host disease (aGvHD) treatment.¹ AREG was identified as a biomarker in a subanalysis of the phase I portion of the urinary-derived human chorionic gonadotropin/epidermal growth factor (uhCG/EGF) study (summarized here) and was found to be high at the start of treatment and diminishing in responders.

Study design and patient characteristics

The analysis of biomarkers used a subset of patients (n = 51) from the current uhCG/EGF trial plus data from a separate cohort of patients treated with ruxolitinib for steroid refractory GvHD as part of the REACH-1 trial (n = 60; NCT02953678), which was used as a validation cohort. As well as AREG, the previously identified potential biomarkers for aGvHD outcomes, suppressor of tumorgenicity 2 (ST2) and regenerating islet-derived 3a (REG3a), were assessed as monitoring biomarkers for aGvHD.

Plasma samples were obtained as shown in Figure 1. In the uhCG/EGF study, AREG was measured using enzyme-linked immunosorbent assay (ELISA), and a bead-based multiplex assay was used to measure ST2 and REG3a. In the REACH-1 trial, a microfluidic immunoassay was used to measure all biomarkers.

*Adapted from Holtan.¹

As shown in Table 1, in the REACH-1 cohort, the median age was 52 years and 52% of patients were male, whereas the uhCG/EGF cohort had a slightly increased median age of 61.5 years and 75% of patients were male. In the REACH-1 trial, patients were fairly evenly split between the Mount Sinai Acute GvHD International Consortium aGvHD grades. The patients in the first-line Minnesota high-risk group (MHR group) in the uhCG/EGF study were mostly Grade II, whereas the second-line group (2LT group) showed a more even distribution among grades.

Table 1. Baseline patient characteristics in both studies*

2LT, second-line treatment; aGvHD, acute graft-versus-host disease; MAGIC, Mount Sinai Acute GvHD International Consortium; MHR, Minnesota high-risk; uhCG/EGF, urinary-derived human chorionic gonadotropin/epidermal growth factor. *Adapted from Holtan.1
Characteristic	uhCG/EGF		REACH-1 (n = 60)
	MHR group (n = 26)	2LT group (n = 26)
Median age (range), years	61 (22−72)	62 (2−69)	52 (18−73)
Male, %	73	77	52
aGvHD clinical grade at enrollment, %
II	8	35	—
III	81	31	—
IV	12	35	—
MAGIC aGvHD grade, %
II	—	—	37
III	—	—	40
IV	—	—	23

Results

At baseline in the uhCG/EGF cohort, AREG levels were not significantly different between patients who went on to achieve complete response (CR), partial response (PR), or no response (NR) at Day 28. However, ST2 and REG3a were significantly elevated for patients with NR vs CR at Day 28 (p < 0.01 and p < 0.05, respectively).

In the REACH-1 study, baseline AREG was increased for patients with progressive disease (PD) at Day 28 compared with those with CR (p < 0.01). For ST2, there was a significant difference in levels at Day 28 for patients with PD compared with CR (p < 0.01), as well as for those with a very good partial response (VGPR) compared with PR (p < 0.05).

There was no significant difference in baseline REG3a in any outcome category at Day 28.

When AREG levels were investigated over the course of the uhCG/EGF trial, a significant decrease was seen between baseline and Day 56 in patients who achieved CR (mean, 98 pg/ml vs 32 pg/ml at baseline and Day 56, respectively; p = 0.006). No significant change was seen in AREG levels for patients who reached PR or had NR.

In the REACH-1 dataset, a significant decrease in AREG levels was observed between baseline and Day 56 in patients that achieved CR on Day 28 (mean, 174.7 pg/ml vs 63.6 pg/ml at baseline and Day 56, respectively; p = 0.007), as well as for those with a PR or VGPR at Day 28 (mean, 288.2 pg/ml vs 146.1 pg/ml at baseline and Day 56, respectively; p = 0.017). No significant difference in AREG level was seen for patients with PD between baseline and Day 56.

Regarding ST2 levels, there was a 1.4-fold decrease between baseline and Day 56 for patients with CR by Day 28 (p = 0.02).

There was no significant change in REG3a levels between baseline and Day 56 in any outcome category.

Biomarker cutoff points associated with a fatal course

For the three biomarkers, cutoff points were identified that were associated with a rapidly fatal disease course. The cutoff points, as shown in Table 2, demonstrate a tendency towards being increased in the REACH-1 study compared with the uhCG/EGF trial. Whether this reflects a difference in how the biomarkers were measured or the severity of the illness is unclear.

Table 2. Biomarker cutoffs across both trials*

AREG, amphiregulin; NR, not reached; REG3a, regenerating islet-derived 3a; ST2, suppressor of tumorgenicity 2; uhCG/EGF, urinary-derived chorionic gonadotrophin/epidermal growth factor. *Data from Holtan.1
Biomarker	UhCG/EGF			REACH-1
	Cutoff	High vs low biomarker levels		Cutoff	High vs low biomarker levels
		Median survival, days	p value		Median survival, days	p value
AREG	212 pg/ml	62 vs NR	0.006	336 pg/ml	74 vs NR	0.005
ST2	292 ng/ml	239 vs NR	<0.001	188 ng/ml	—	0.09
REG3a	13.5 ng/ml	416 vs NR	0.01	3.6 ng/ml	—	0.3

The only factors that were shown to be independently associated with survival in multivariate analysis were response at Day 28 (NR vs PR/CR; risk ratio, 4.94; p = 0.02) and baseline AREG >212 pg/ml (risk ratio, 4.17; p = 0.03). This was similar for the REACH1 study, with Day 28 response (PD vs CR/VGPR/PR; risk ratio, 9.14; p < 0.0001) and baseline AREG >336 pg/ml (risk ratio, 2.72; p < 0.05) being associated with survival.

Conclusion

The use of AREG shows promise as a monitoring biomarker for patients with aGvHD, with high levels of baseline AREG being associated with an increased risk of early mortality. These findings were replicated using samples from two different studies utilizing different biomarker assessment platforms. Holtan concluded that patients with high baseline AREG should be closely monitored using serial assessments throughout treatment.