The HB22.7–vcMMAE antibody–drug conjugate has efficacy against non-Hodgkin lymphoma mouse xenografts with minimal systemic toxicity

Abstract

In this study, HB22.7, an anti-CD22 monoclonal antibody, was used for specific, targeted delivery of monomethyl auristatin E (MMAE) to non-Hodgkin lymphoma (NHL). MMAE was covalently coupled to HB22.7 through a valine–citrulline peptide linker (vc). Maleimide-functionalized vcMMAE (mal-vcMMAE) was reacted with thiols of the partially reduced mAb. Approximately 4 molecules of MMAE were conjugated to HB22.7 as determined by residual thiol measurement and hydrophobic interaction chromatography–HPLC (HIC-HPLC). HB22.7–vcMMAE antibody–drug conjugate (ADC) retained its binding to Ramos NHL cells and also exhibited potent and specific in vitro cytotoxicity on a panel of B cell NHL cell lines with IC₅₀s of 20–284 ng/ml. HB22.7–vcMMAE also showed potent efficacy in vivo against established NHL xenografts using the DoHH2 and Granta 519 cell lines. One dose of the ADC induced complete and persistent response in all DoHH2 xenografts and 90 % of Granta xenografts. Minimal toxicity was observed. In summary, HB22.7–vcMMAE is an effective ADC that should be evaluated for clinical translation.

Electronic supplementary material

The online version of this article (doi:10.1007/s00262-016-1873-y) contains supplementary material, which is available to authorized users.

Introduction

In 2014, approximately 70,800 people in the USA alone were diagnosed with non-Hodgkin lymphoma (NHL) [1]. The overall 5- and 10-year survival rates are 71 and 63 %, respectively [2]. The most common current standard of care for B cell NHL is chemotherapy combined with the anti-CD20 monoclonal antibody (mAb), rituximab. Commonly used regimens include rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), bendamustine and rituximab, rituximab, cyclophosphamide and prednisone, and fludarabine-based combinations. These regimens have many off-target effects that result in substantial toxicity that limit efficacy. The median age at diagnosis for NHL is 65, which highlights the importance of reducing treatment-related toxicity. Monoclonal antibodies may be used as targeting molecules with a goal of reducing systemic toxicity while enhancing efficacy by the direct delivery of toxins specifically to cancer cells [3]. Therefore, antibody–drug conjugates (ADCs) may not only potentiate current chemotherapeutics but also permit lower effective doses, thus reducing the toxicity of systemic chemotherapy.

The CD22 receptor, a member of the SIGLEC family, is expressed by normal and malignant B cells and has been shown to be involved in the regulation of B cell function and survival [4]. CD22 is rapidly internalized upon binding of the mAb HB22.7, making it an ideal target for ADC therapy due to efficient intracellular delivery of conjugated payloads [5–7].

HB22.7 is an anti-CD22 mAb originally developed to map the CD22 ligand-binding domain [8]. Based on its ligand-blocking properties, HB22.7 was humanized to aid in the treatment of NHL. The CD22 ligand-binding domain mediates a B cell survival signal [9] while anti-CD22 antibodies that block CD22 ligand-binding induce apoptosis in B cell malignancies [10–12]. We have previously compared the in vitro and in vivo activity of CD22 ligand-blocking versus non-blocking antibodies and found that ligand-blocking antibodies posses superior cytotoxicity [12]. In addition to its apoptosis-inducing properties, we have also shown that HB22.7 can be an effective targeting vehicle that mediates cellular uptake of conjugated payloads [3].

In the present study, we examined the potential of HB22.7 to mediate targeted delivery of a directly conjugated drug, monomethyl auristatin E (MMAE, a potent antimitotic drug that inhibits cell division and induces apoptosis by binding to microtubules and inhibiting polymerization [13, 14]). Several groups have developed mAb-MMAE ADCs to a variety targets, including CD19, CD22, CD79b, and CD30, AGS-5, and guanylyl cyclase C [15–18]. Most have demonstrated significant preclinical and clinical efficacy. Moreover, a CD30-targeted MMAE ADC was approved by the FDA in 2012 for the treatment of relapsed Hodgkin and anaplastic lymphomas and has proven to be an effective and well-tolerated treatment option. Pinatuzumab vedotin (DCDT2980S), an anti-CD22 MMAE conjugate developed by Genentech, is also under clinical investigations for the treatment of relapsed or refractory diffuse large B cell lymphoma and follicular NHL in combination with rituximab [16].

The independent lymphomacidal activity of HB22.7 due to its ligand-blocking properties (in contrast to other anti-CD22 mAbs that do not block ligand-binding) rationalizes the assessment of the HB22.7–MMAE ADC. Herein, we present in vitro and in vivo data using the ligand-blocking monoclonal antibody conjugated to MMAE via a valine–citrulline (vc) peptide linker (HB22.7–vcMMAE) that demonstrates significant preclinical efficacy.

Materials and methods

Materials

Dithiothreitol (DTT) was obtained from Acros Organics. 5,5′-dithiobis-2-nitrobenzoic acid (DTNB) was purchased from Thermo Scientific. Diethylenetriaminepentaacetic acid (DTPA) was purchased from Sigma-Aldrich. Maleimide-vc-MMAE (mal-vcMMAE) was a gift from Dr. Zhenwei Mao (Concortis Biosystems). HB22.7 was prepared and characterized as previously described [11]. Anti-CD22 sc7323 was from Santa Cruz Biotechnology.

Cell lines

The lymphoma cell lines Ramos, Raji (Burkitt’s lymphoma) Granta 519 (mantle cell lymphoma), and the T cell leukemia cell line Jurkat were purchased from the American Type Culture Collection. The lymphoma cell lines SU-DHL-4 (diffuse large B cell lymphoma) and DoHH2 (human transformed follicular lymphoma) were purchased from DSMZ. All cell lines used in this study were maintained in RPMI 1640 with 10 % fetal bovine serum at 37 °C, 5 % CO₂, and 90 % humidity.

Synthesis of HB22.7–vcMMAE

HB22.7–vcMMAE was prepared using a limited DTT reduction strategy as previously described [19]. Briefly, HB22.7 (10–20 mg/ml in PBS) was incubated with 3.25 molar equivalents of DTT for 2 h at 37 °C. Excess DTT was removed from the partially reduced HB22.7 by passing the mixture over a PD-10 column and eluted with PBS. Fractions were collected and assessed at A₂₈₀ to determine the protein-containing fractions. Fractions containing thiolated HB22.7 were then concentrated with YM-30 Centricon ultrafiltration devices. The final HB22.7 concentration was determined using the A₂₈₀ and a molar extinction coefficient of 1.35. The ratio of thiols per mAb was determined by mixing thiolated HB22.7 with 0.1 mM DTNB (Ellman’s reagent) and measuring at A₄₁₂ with a molar extinction coefficient of 13,600 M⁻¹. DTPA (1 mM) was added to the reduced HB22.7 to prevent oxidation of thiols.

A stock solution of mal-vcMMAE was prepared in 50 % acetonitrile. Partially reduced HB22.7 was mixed with up to 1.5 equivalents of mal-vcMMAE with a final acetonitrile concentration of 5 % to ensure solubility. The reaction proceeded for 2 h at 4 °C. The degree of drug loading was calculated by quantitation of residual thiols after drug conjugation or by hydrophobic interaction chromatography–high-performance liquid chromatography (HIC-HPLC) [14, 19].

Flow cytometry

CD22 expression and HB22.7–vcMMAE binding were assessed by flow cytometry. Ramos cells (0.5 × 10⁶ per sample) were resuspended in 100 µl FACS buffer (PBS/0.5 % FBS) and incubated with HB22.7 (10 µg/ml) at 4 °C for 30 min with “head over heels” rotation. Cells were washed 3× and further incubated for 30 min with a 1:50 dilution of FITC-labeled secondary antibody (goat antihuman IgG, Invitrogen). After 3 washes, cells were resuspended in FACS buffer and analyzed using a FACScan instrument (BD Biosciences). Ten thousand events per sample were acquired.

Cytotoxicity

In vitro cytotoxicity of HB22.7–vcMMAE was evaluated using an MTS assay. Briefly, 10⁴ cells/90 μl/well in 96-well plates were incubated with 10 µl of serial dilutions of HB22.7–vcMMAE for 72 h at 37 °C, 5 % CO₂. A CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega) was used to assess viability according to the manufacturer’s instructions. Cell viability was calculated by OD₄₉₀ measurements and normalized to the untreated control. Three independent experiments, each in triplicates, were performed; data are presented as the mean ± standard deviation (SD).

NHL xenograft experiments

Female (6–8 weeks old) ICR-SCID mice were purchased from Charles River and maintained in micro-isolation cages under pathogen-free conditions. After an acclimatization period of at least 4 days, animals were irradiated (400 rads, whole body). Xenografts were established 3 days after irradiation by subcutaneous flank injections of 10⁷ Granta 519 or DoHH2 cells in 100 µl PBS. Once tumors reached 100–200 mm³ (designated as day 0), mice were randomized into two groups (8–10 per group) consisting of untreated control or HB22.7–vcMMAE (7.5 mg/kg). The dose of HB22.7–vcMMAE was based on previous studies with similar ADCs in similar animal models [11–13]. Treatments were administered on day 0 by intraperitoneal (ip) injection. Digital calipers were used to take measurements, and tumor volume was calculated as (L × W ²)/2, where L and W are, respectively, the longer and shorter dimensions of the tumor.

Body weight and signs of toxicity were recorded twice a week. Mice were euthanized when tumor volume exceeded 1500 mm³ or 20 mm in. in any dimension according to the institutional regulations.

All animal experiments were in compliance with Institutional (UC Davis IACUC) and Federal guidelines under approved animal protocols.

Results

Development of HB22.7–vcMMAE

The conjugation of vcMMAE to HB22.7 was done as previously described with modifications as described in “Materials and methods” section. The ADC was produced in house, and the drug-to-antibody ratio (DAR) of 4.6 was calculated by a third party (Concortis) via HIC-HPLC (see Supplementary Fig. 1).

Flow cytometric analysis of HB22.7–vcMMAE binding

To ensure that MMAE conjugation did not affect the antibody binding to its target, the CD22+ cell line Ramos was used to compare the binding of the parent (HB22.7) to that of HB22.7–vcMMAE. MMAE conjugation to HB22.7 had no effect on its binding to Ramos cells (Fig. 1).

Fig. 1.

Flow cytometric analysis of HB22.7–vcMMAE. Binding of HB22.7 to CD22 after MMAE conjugation. Ramos cells were incubated with 10 µg/ml of either HB22.7 alone (black) or HB22.7–vcMMAE conjugate (blue). Secondary antibody only was used as control (solid red)

HB22.7–vcMMAE and free MMAE in vitro cytotoxicity

We used a panel of CD22+ NHL cell lines representing NHL subtypes to assess the in vitro cytotoxicity of HB22.7–vcMMAE. As shown in Fig. 2a, all cell lines tested were sensitive to free MMAE with IC₅₀s ranging from 0.099 to 1.348 nM. HB22.7–vcMMAE was effective in all CD22+ NHL cell lines tested, Fig. 2b. IC₅₀s for the cell lines ranged from 0.02 to 0.285 µg/ml, and while all cell lines expressed CD22, the degree of cytotoxicity did not correlate with CD22 expression levels (see Supplementary Figs. 2–3). DoHH2 cells were the most sensitive with an IC₅₀ of 0.02 µg/ml. Although the CD22-negative cell line Jurkat was very sensitive to free MMAE with an IC₅₀ of 0.099 nM, it was, as expected, resistant to HB22.7–vcMMAE with an IC₅₀ greater than 2.5 µg/ml, as shown in Fig. 2 and Table 1. In Jurkat cells, the lack of killing by the conjugate, while unconjugated MMAE was toxic, suggests that the conjugate is CD22-specific and stable since it did not release free MMAE into the culture media. Treatment with the naked antibody (HB22.7, isotype control or the commercially available sc7323) at similar concentrations did not result in detectable cytotoxicity.

Fig. 2.

In vitro cytotoxicity and selectivity of HB22.7–vcMMAE. NHL cell lines were treated with escalating doses of either free HB22.7 (a) or HB22.7–vcMMAE conjugate (b). Cytotoxicity was evaluated by MTS assays after 72 h of continuous exposure. Viability is shown as relative to the untreated controls. Error bars indicate SD (n = 3)

Table 1.

IC₅₀ values for free MMAE, free antibody, and HB22.7–vcMMAE on NHL cell lines

Cell line	IC50
	MMAE (nM)	HB22.7–vcMMAE (µg/ml)	HB22.7 (µg/ml)	sc7323 (µg/ml)	Isotype (µg/ml)
Ramos	0.104	0.038	>20	>20	>20
Raji	0.705	0.284
SU-DHL-4	0.292	0.192
Granta 519	1.348	0.070
DOHH2	0.289	0.020
Jurkat	0.099	>2.5

In vivo efficacy of HB22.7–vcMMAE

The preclinical efficacy of HB22.7–vcMMAE was evaluated using Granta 519 and DoHH2 cell lines. Xenograft tumors were established in the flank of ICR-SCID mice, and treatment was administered once tumors reached approximately 100 and 170 mm³ for DoHH2 and Grant 519, respectively. In untreated mice, tumors reached 1500 mm³ within 15–20 days. In comparison, treatment with one dose of HB22.7–vcMMAE (7.5 mg/kg) induced complete, durable tumor remission in all mice bearing DoHH2 xenografts (Fig. 3a) and in 9 out of 10 mice bearing Granta xenografts (Fig. 3b). Body weight and complete blood counts (CBC) were used to assess treatment-induced toxicity. Average percent of starting body weights in the treated groups were not significantly different from controls (Fig. 4a, b). Mice treated with HB22.7–vcMMAE had CBCs within the normal range, and the blood counts did not deteriorate over time (Fig. 4c–e).

Fig. 3.

Efficacy of HB22.7–vcMMAE in transformed follicular and mantle cell lymphoma xenograft models. Flank NHL xenografts were established with DoHH2 (a) or Granta (b) cells. Mice (5–10/group) were treated with 7.5 mg/kg of HB22.7–vcMMAE administered as a single dose at day 0 (black) when the average tumor size reached 100 (DoHH2) and 170 mm³ (Granta 519). Control mice (5/group) were monitored without treatment (blue) until the tumor size reached 1500 mm³ at which point they were euthanized

Fig. 4.

No signs of toxicity were observed in mice treated with HB22.7–vcMMAE compared to untreated control. No toxicity was associated with the treatment, as assessed by body weight (a, b) and complete blood counts (c–e)

Discussion

In previous studies, our group used HB22.7 as a vehicle for targeted delivery of saporin (SAP) to CD22+ NHL. SAP was chosen as the cytotoxic payload because of its ease of chemical conjugation to antibodies and its potency. Although the HB22.7–saporin conjugate was efficient at targeting CD22+ malignancies, it did not induce complete regression in human NHL xenograft tumor models. Moreover, clinical development of a mAb-SAP conjugate may be challenging due to hepatotoxicity and the development of neutralizing anti-SAP antibodies [20]. Preclinical efficacy of antibody-mediated targeted delivery of SAP has been reported by several groups [21, 22]. However, toxins produced by plants, fungi, and bacteria are easily recognized as foreign molecules by the immune system leading to the development of neutralizing antibodies which in turn results in rapid clearance of the conjugate and hence reduced levels available for uptake by tumor cells.

As an alternative to SAP, potent small molecule drugs such as those derived from the maytansine (DM1), auristatin, and monomethyl auristatin E (MMAE) families have shown promising results as the cytotoxic payload of ADCs. These drugs have similar or higher potency compared to SAP but have the distinct advantage of not producing the above-mentioned side effects. Several groups have successfully conjugated MMAE to mAb and showed promising preclinical and clinical efficacy. Recently, Genentech has developed anti-CD22 antibody and MMAE conjugate (DCDT2980S) which has shown promising efficacy against NHL xenograft models. DCDT2980S, pinatuzumab vedotin (PiV), is currently undergoing Phase II clinical trials in combination with rituximab in patients with relapsed or refractory B cell NHL [16, 23].

The independent lymphomacidal activity and successful utilization of HB22.7 to target CD22-positive B cell malignancies led us to further investigate the use of a more potent and non-immunogenic drug such as MMAE [3] in conjunction with an anti-CD22 mAb that has ligand-blocking and demonstrated independent lymphomacidal activity. HB22.7–vcMMAE consists of the anti-CD22 mAb, HB22.7, chemically conjugated to MMAE via a stable biodegradable peptide linker composed of valine and citrulline. MMAE is a synthetic antimitotic agent that inhibits cell division by blocking the polymerization of tubulin [24]. The linker between HB22.7 and MMAE is stable in extracellular fluid, but is cleaved by cathepsin once the conjugate has entered tumor cells, thus activating the antimitotic activity of MMAE intracellularly [25].

All the cell lines tested in this study were sensitive to free MMAE, with IC₅₀ values lower than 1.3 nM. Our in vitro results show that, when conjugated to HB22.7, MMAE selectively kills CD22+ NHL cell lines but not the CD22-negative cell line Jurkat. While the toxicity of this ADC is dependent on CD22 expression, and suggests that the delivery of the drug is indeed dependent on antibody targeting, the degree of cytotoxicity did not correlate with CD22 expression levels. The lack of correlation between CD22 expression levels and cytotoxicity has been observed with other MMAE constructs [14] and may suggest that the high potency of MMAE allows for lower target threshold levels. To study the in vivo efficacy of HB22.7–vcMMAE, we treated two xenograft models that represent particularly resistant subtypes of human NHL: transformed follicular and mantle cell lymphoma. In both models, a single dose of the HB22.7–vcMMAE ADC induced durable complete response in nearly all mice (Fig. 3). These results suggest that HB22.7–vcMMAE has potent antitumor activity in vivo with no observable toxicity (Fig. 4).

In summary, our results indicate that the HB22.7–vcMMAE ADC has potent preclinical efficacy against two very difficult-to-treat NHL, mantle cell and transformed follicular lymphoma. Nearly all tumors completely responded to only one dose of HB22.7–vcMMAE. The results from this study are very comparable to Genentech’s ADC, PiV, and other anti-CD22 MMAE ADCs that have been translated to clinical studies.

Previous studies have demonstrated that the ligand-binding domain of CD22 mediates B cell survival, and that blocking this region leads to B cell apoptosis and independent lymphomacidal activity [10, 12]. In addition, because we have previously shown that HB22.7 has independent lymphomacidal activity in vivo [26], it is tempting to hypothesize that the ligand-blocking properties of HB22.7–vcMMAE may contribute to additional cytotoxicity and efficacy. Detailed mechanistic studies to test this hypothesis are the focus of ongoing efforts. While many of the currently anti-CD22 mAbs in clinical development bind to CD22 extracytoplasmic domains that do not block ligand-binding (LL2 and RFB4), the CD22-binding domain targeted by the mAB used in the development of pinatuzumab (10F4) has not been described.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ADC

Antibody–drug conjugate

CBC

Complete blood counts

DAR

Drug–antibody ratio

DTNB

5,5′-dithiobis-2-nitrobenzoic acid

DTPA

Diethylenetriaminepentaacetic

DTT

Dithiothreitol

FDA

Food and drug administration

HIC-HPLC

Hydrophobic interaction chromatography–high-performance liquid chromatography

ip

Intraperitoneal

mAb

Monoclonal antibody

Mal-vcMMAE

Maleimide-functionalized vcMMAE

MMAE

Monomethyl auristatin E

NHL

Non-Hodgkin lymphoma

PiV

Pinatuzumab vedotin

R-CHOP

Rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone regimen

SAP

Saporin

SD

Standard deviation

vc

Valine–citrulline peptide linker

Compliance with ethical standards

Conflict of interest

All authors declare no conflict of interest.