Introduction
Globally, an estimated 270,000 new cases and 116,000 deaths attributed to kidney cancer occur each year [
1]. Greater than 90 % of kidney neoplasms are classified as renal cell carcinoma (RCC), of which upwards of 30 % progress towards metastatic disease [
2,
3]. Treatment of RCC can be broadly grouped into either immunotherapy or targeted therapy. Immunotherapy with high dose interleukin-2 (IL-2) alone or in combination with interferon-α (IFN-α) has historically been a frontline of defense against RCC [
4]. More recently, checkpoint control inhibitors and targeted therapies have shown great promise in combating RCC [
2,
5,
6]. The currently approved targeted therapies against RCC include a monoclonal antibody (mAb) that blocks vascular endothelial growth factor from binding its receptor (bevacizumab), tyrosine kinase inhibitors (TKIs; sunitinib, sorafenib, pazopanib, and axitinib), and mammalian target of rapamycin inhibitors (everolimus and temsirolimus). These therapies alone or in combination have shown promise towards combating RCC [
7,
8], but a complete and curative regimen by these drug or immunotherapy agents remains elusive.
Tumor-associated antigens are the focus of both diagnostic and therapeutic strategies against many forms of cancer, and carbonic anhydrase (CA) IX is the most well characterized antigen associated with RCC [
9,
10]. CAIX is a trans-membrane protein that alters the pH of the extracellular microenvironment towards acidity during hypoxia [
10‐
12]. Expression of CAIX is controlled by the tumor suppressor von Hippel-Lindau factor, and nearly all forms of clear cell RCC express CAIX, which is absent on normal kidney tissue [
9]. Expression of CAIX on RCC is a prognostic factor, in that reduced CAIX expression is closely associated with the progression of disease [
13] while expression of CAIX correlates with the efficacy of some treatments [
4]. Thus, although expression of CAIX may provide RCC tumors with a growth advantage [
14‐
16], it also conveys a level of susceptibility to treatment.
Multiple immunotherapeutic approaches center on targeting CAIX, including cellular-based strategies to vaccinate against CAIX [
17‐
19] and mAbs that bind CAIX [
20‐
23]. The chimeric antibody girentuximab (Rencarex®, Wilex) has been extensively tested in clinical trials [
24], including its use alone, coupled to radioisotopes, in combination with cytokines, and in conjunction with chimeric antigen receptor (CAR) T cell therapy [
25]. Results were recently released from a Phase 3 trial testing the efficacy of girentuximab alone versus placebo to limit disease progression in patients having undergone nephrectomy of non-metastatic clear cell RCC [
26]. While the study objectives were not met, preliminary data indicate that a higher degree of efficacy was seen in younger patients whose tumors expressed high levels of CAIX. These findings, in conjunction with other clinical and preclinical data, emphasize the potential for anti-CAIX mAbs in treating RCC, but also highlight the need for new anti-CAIX mAbs that may have more potent anti-tumor properties [
2,
27].
In seeking to develop novel therapeutic agents against RCC and other CAIX-expressing malignancies, we sought to further characterize a panel of high affinity, fully human antibodies against CAIX that we had previously described [
23]. We focused on two high-affinity antibodies that inhibit CA activity, but have disparate capabilities to induce CAIX internalization. We examined their ability to elicit cellular immune responses to CAIX expressing tumor cells both
in vitro and
in vivo. Our results demonstrate that these two human anti-CAIX mAbs mediate immune killing of CAIX
+ tumor cells
in vitro and show potent therapeutic activity
in vivo.
Discussion
Although treatment of RCC with antibody-based therapies has increased over the last several years [
5], cancer cures with this approach has remained elusive [
2,
3]. Various antibodies targeting the tumor-associated antigen CAIX on RCC cells have been amongst those most extensively studied [
8,
21,
23,
24,
34], owing to the antigen’s high expression on RCC. One chimeric anti-CAIX antibody that does not inhibit enzymatic activity, girentuximab, has been evaluated in the clinic [
24,
43]. In preliminary analysis of Phase 3 trial results, an increase in the duration of disease-free survival was seen only in a subgroup of younger patients who exhibited high expression of CAIX on RCC tumors. Several small molecules known to effectively inhibit CA have demonstrated particular promise as potential anti-cancer agents [
44,
45]. These findings highlight a need to both more clearly define patient populations under investigational drug study and develop novel anti-CAIX antibody drug candidates to increase overall response rate of patients with CAIX
+ RCC tumors. Thus, anti-CAIX antibodies that combine inhibition of the enzymatic activity of CAIX with ADCC or CDC might be more potent than girentuximab.
In this study, we selected several of our anti-CAIX mAbs for further characterization based on their high affinity, capacity to inhibit CA activity, or ability to induce CAIX internalization (Additional file
6: Table S1) [
23]. Our findings indicate that when CAIX-expressing RCC cells were simultaneously incubated with internalizing or non-internalizing anti-CAIX mAbs and PBMCs, ADCC activity of the antibodies showed no correlation with their reported capacity to induce CAIX internalization (Fig.
1). Similarly, the capacity to induce CAIX internalization did not affect the ability of anti-CAIX mAbs to limit RCC motility, as both G37 (non-internalizing) and G119 (internalizing) IgG1 demonstrated comparable inhibition of RCC expansion and movement
in vitro (Fig.
2). In addition, Fc engineering to enhance effector activity [
31] increased the ADCC mediated killing of RCC by both mAbs (Fig.
3a), but had only marginal effects on CDC and phagocytosis (Figs.
3b and c). This enhanced
in vitro ADCC killing activity did not translate into more potent inhibition of tumor growth
in vivo. However, this may have occurred because of the high doses (10 mg/kg) and frequent treatments (2x/week) of the anti-CAIX mAbs. Additional dose response studies on the wild type and Fc mutated mAbs should be performed in future studies to determine if an enhancement in potency of tumor killing activity by Fc engineered antibodies could be seen at lower concentrations
in vivo.
Animal models examining the efficacy of cancer therapeutics increasingly use a xenograft approach where cancer cell lines or patient derived tumor tissue are implanted into immunodeficient mice [
27]. These patient-derived tumor models (PDTM) have been used in several studies of RCC [
33,
34,
46,
47], and offer a promising means by which targeted therapeutics such as tyrosine kinase inhibitors can be evaluated. A caveat to this approach, however, is that the complexities of tumor and immune cell interactions are not evaluated in current RCC PDTM, and therefore these critical parameters are often missing or underrepresented in the analysis [
33]. This shortcoming is mirrored in the analysis of antibody therapeutics, most immediately because of a lack of a functional immune system in mice engrafted with human tumors. We sought to circumvent this by developing an allogeneic mouse model whereby the capacity of antibody therapies against CAIX can be evaluated (Fig.
4). To this end, we created a model wherein human donor PBMC with a demonstrated capacity to kill CAIX
+ RCC cells by ADCC (Additional file
2: Figure S2) were injected into mice bearing an orthotopic RCC tumor. To characterize the tumor killing activity mediated by anti-CAIX mAbs, a luciferase-expressing CAIX
+ RCC cell line was transplanted orthotopically to facilitate measurement of tumor growth. We anticipated that these efforts would afford us the capacity to examine the efficacy of anti-CAIX mAbs in the presence of human immune cells to inhibit tumor growth
in vivo.
While our mouse model demonstrated that both wild type and mutant Fc formats of anti-CAIX G37 and G119 IgG1 were capable of limiting tumor growth, this observation did not manifest by BLI until circa three weeks after PBMC and anti-CAIX treatment (Fig.
6). A possible reason for the delay in detecting tumor inhibition may be the limitations of BLI sensitivity in measuring small orthotopic tumors. Alternatively, tumor inhibition may have been delayed because several subsets of immune cells in human PBMCs do not persist in the NSG mouse for long periods of time.
Consistent with previous reports [
48,
49], we were unable to detect human NK cells in the mice 13 days after PBMC injection (Fig.
7b). However, we did find that a greater influx of human NK cells occurred in tumors one day after antibody injection (7 days after PBMC injection) in mice treated with anti-CAIX mAbs (Fig.
7a). This indicates that, while human NK cells do not persist, these cells likely contributed to early inhibition of tumor growth by the anti-CAIX mAbs, which did not manifest overtly until much later time points.
An additional possibility for the inhibition of tumor growth seen in anti-CAIX mAb-treated mice may involve the influence and interaction of human T cells and mouse myeloid cells. Human T cells infiltrating the engrafted tumor appear most activated at later time points in our model, as measured by expression of human PD-1 and IFN-γ [
42] (Fig.
8 and Additional file
4: Figure S4). This activation of human T cells is coincident with mouse myeloid influx in the tumor. While the myeloid compartment in NSG mice has been characterized as deficient [
50], it is possible that shed CAIX/anti-CAIX complexes can active the mouse myeloid compartment and through direct cellular interactions and cytokine secretion lead to further activation of human T cells to express PD-1 (Fig.
8b) [
51]. It is also possible that graft versus host disease [
52] or HLA-mismatch with the engrafted tumor can act in concert with anti-CAIX mAbs and mouse myeloid cells to enhance CAIX
+ tumor cell killing. It should also be noted that the SKRC-59 cells used for the
in vivo studies herein are CAIX
+PD-L1
+ (data not shown) and therefore the potential role that the PD1:PD-L1 checkpoint control axis may play in limiting the Teff response must also be considered.
Finally, PBMC from the terminal time point (including both human immune and mouse myeloid cells) demonstrate ADCC activity, despite lacking human NK cells, which suggests that enhancement of NSG myeloid cell activity may have occurred (Fig.
7c). While our data demonstrates a lack of CDC activity
in vivo (Additional file
3: Figure S3b) owing to a 2-base pair deletion in the complement C5 structural gene [
53], it is possible that anti-CAIX deposition on the tumor cells can lead to activation of earlier complement components including C3a and C4a that can lead to chemotactic recruitment of myeloid cells and C3b to promote opsonization. Both of these actions could lead to myeloid cell activation and enhanced tumor killing activity [
54]. This speculation clearly requires more extensive characterization of infiltrating human T cell and mouse myeloid cell activity in the tumor. Therapeutic antibodies targeting the PD-1:PD-L1 axis [
55,
56] in combination with anti-CAIX mAbs would be particularly amenable to assessment with our model as PD-L1 is known to be expressed on RCC cells and associated with poor outcome [
57].
Materials and methods
Cell lines and culture
Human renal clear cell carcinoma (RCC) cell lines, SKRC-09 (CAIX positive), SKRC-52 (CAIX positive), and SKRC-59 (CAIX negative), were obtained from Dr. Gerd Ritter (Memorial Sloan-Kettering Cancer Center, New York, NY). These lines were grown in RPMI 1640 Medium (Life Technologies) supplemented with 10 % (v/v) heat-inactivated fetal bovine serum (FBS, Gibco), 100 IU/ml penicillin and 100 μg/ml streptomycin (complete medium) at 37 °C with 5 % CO2. RCC4 and 293 T cells (CRL-11268, ATCC) were grown in DMEM Medium (Life Technologies) supplemented with 10 % FBS, 100 IU/ml penicillin and 100 μg/ml streptomycin. The 293 F cell line (Life Technologies) was propagated and transfections for protein production carried out in 293 FreeStyle serum-free medium (Life Technologies).
Antibodies and flow cytometry analysis
Anti-CAIX antibody variants (G10, G36, G37, G39, and G119) and control antibody were produced as described previously [
23,
30]. Briefly, scFv-Fcs were constructed by cloning the single-chain variable region (scFv) in frame with a human IgG1 Fc region lacking a CH1 domain. Full length, human IgG1s were generated by cloning heavy chain variable region (VH) and light chain variable region (VL) into a dual-promoter expression cassette [
4]. Antibodies (scFv-Fc and IgG1) were transfected into 293 F cells using polyethylenimine and purified by protein A sepharose affinity chromatography via ÄKTAPurifier FPLC (GE Healthcare). To characterize nature killer cells, cells were stained with human CD45, CD56 and CD16 antibodies (BioLegend) and analyzed by flow cytometry.
Antibody-dependent cell-mediated cytotoxicity assay
Cytotoxicity assays utilized measurement of lactose dehydrogenase (LDH) release, with PBMC from healthy human donors or NSG mice used as effector cells. RCC cell lines SKRC-09, SKRC-52, SKRC-59, or CAIX
+ SKRC-59 (engineered to express CAIX by lentiviral transduction [
23]) were used as target cells and plated at 1 × 10
4/well in a 96-well plate. Cells were incubated with indicated concentrations of antibody for one hour, followed by co-culture with effector cells for 4–6 h [
4]. Culture supernatants were harvested by centrifugation and LDH measured in the supernatant by colorimetric assay (Promega) at 490 nm. The percent cytotoxicity was calculated as: %Cytotoxicity = 100 × (E – SE – ST)/(M – ST); E, released LDH from E/T culture with antibody; SE, spontaneous released LDH from effectors; ST, spontaneous released LDH from targets; M, the maximum released LDH from lysed targets. All tests were performed in triplicate.
Complement dependent cytotoxicity assay
SKRC-52 cells were cultured in 96 well plates (2 × 104/well) RPMI 1640 Medium (Life Technologies), supplemented with 100 IU/ml penicillin and 100 μg/ml streptomycin, and rabbit serum (Cedarlane Laboratories). Control or CAIX antibodies were added at the indicated amounts and cytotoxicity measured as in ADCC assays by LDH release after 6 h. Cytotoxicity was calculated as: %Cytotoxicity = 100 × (E – ST)/(M – ST). All tests were performed in triplicate.
Migration assay
Migration of SKRC-52 cells was determined by transwell assay using FluoroBlok 24-multiwell insert assay (BD Biosciences). Overnight starved (0.5 % FBS) SKRC-52 cells were labeled with a lipophilic fluorescent dye DiO (Life Technology) and seeded in the upper chamber at 1 × 105 cells/well in 0.5 % FBS RPMI in the presence or absence of 2.5 μg/ml G37, G119, a non-specific IgG1, or 100 μM of acetazolamide as a positive control for inhibition of migration. Hepatocyte growth factor (HGF, Sigma) was added into the lower chamber. Fluorescence intensity of the lower chamber, indicating cell migration, was measured after 24-h incubation at 37 °C using a POLARstar reader (BMG Labtech). All tests were performed in triplicate.
Wound healing assay
SKRC-52 cells were cultured to confluency in 24-well plates (approximately 2 × 105 cells/well). The monolayers were incubated in the absence of serum for 16 h and wounded in a line across the well with a 200-μl pipette tip. The wounded monolayers were washed twice with serum-free media and incubated with 10 μg/ml of G37, G119, a non-specific IgG1, or 100 μM acetazolamide. The area of cell-free wound was recorded 24-h later using a charge-coupled device camera (C2400; NEC) on an inverted microscope (Axiovert 35; Zeiss) and analyzed with image analysis software (NIH Image 1.55). Wound healing was calculated as the percentage of the remaining cell-free area compared with the area of the initial wound. All tests were performed in triplicate.
Cell growth inhibition assay
SKRC-52 cells were seeded in 96-well plates at a density of 4 × 103 cells/well. After 24 h, cells were incubated with anti-CAIX antibodies or control antibody in fresh medium. Followed by 24, 48, 72, 96, and 120 h incubation, MTT (2 mg/ml in RPMI, 50 μl/well) was added to the wells and further incubated at 37 °C for 2.5 h. The supernatant was removed and 150 μl DMSO per well was added to dissolve the produced formazan. After shaking the plates for 10 min, optical absorbance at 570 nm were recorded with a microplate reader.
Antibody dependent cellular phagocytosis assay
CD14 positive monocytes were isolated from healthy donor whole blood by fluorescent activated cell sorting (FACS), and cultured in 10 cm2 dishes with complete medium supplemented with 200 U/ml granulocyte macrophage colony-stimulating factor (GM-CSF, eBioscience). Culture medium was refreshed every other day for 5 days, non-adherent cells were washed away with PBS and the adherent monocyte-derived macrophages (MDMs) harvested. SKRC-52 cells were labeled with PKH-26 fluorescent membrane dye (Sigma), washed three times, incubated on ice for 1 h without antibody or CAIX antibodies, and cultured with MDMs at a 1:8 ratio for 4 h at 37 °C. Phagocytosis was determined after extensive washing of the co-culture, and measured by FACS as CD14+PKH+ cells. The percentage of phagocytosis was calculated as: %Phagocytosis = 100 × [(percent dual positive)/(percent dual positive + percent residual targets)]. All tests were performed in triplicate.
Orthotopic RCC tumor: allogeneic human PBMC model
Orthotopic RCC tumors were established by injecting 5 × 104 SKRC-59 cells into the left subrenal capsule of kidneys of NSG mice. SKRC-59 cells were engineered to express high levels of human CAIX and luciferase, through lentiviral transduction, and passaged subcutaneously in NSG mice prior to engraftment to enrich for rapidly growing, CAIX+ cells (data not shown). Prior to injection, RCC cells were suspended in PBS and diluted 1:1 in Matrigel (Life Technologies) to ensure retention of the cells within the subrenal capsule. Three days after engraftment, tumor implantation was confirmed by bioluminescence imaging (BLI) using a Xenogen imaging system (Life Technology) with subsequent measurements taken every 3–4 days thereafter. Four days after tumor engraftment, 1×107 PBMC from a healthy human donor (selected for high ADCC activity) were injected intravenously. Six days after PBMC injection (10 days after engraftment) control or CAIX antibodies were injected at 10 mg/kg intravenously, and subsequently injected every 3–4 days thereafter. Mice were sacrificed and tissues harvested at 32 days post tumor engraftment. Animal care was carried out in accordance with the guidelines of Animal Care and Use Committee of Dana-Farber Cancer Institute.
Histology
RCC tumors and kidneys were surgically removed, fixed in 10 % neutral buffered formalin, embedded in paraffin, and 6-μm thick sections placed on slides. Hemotoxylin and eosin (H&E) immunostaining was performed using anti-human CD3 (A0452, DAKO), anti-human PD-1 (315 M–96, Cell Marque), anti-mouse Ly6G (1A8, BioLegend), anti-IFN-γ (MD-1, BioLegend), and anti-CD8 (M7103, DAKO) antibodies. HRP-labeled anti-rabbit antibodies (Pierce) and Mouse on Mouse (M.O.M.) Kits (Vector) were used as secondary antibodies and detected by diaminobenzidine according to the manufacturer’s instructions. Imaging of sectioned tissues was carried out using the ScanScope system (Aperio) and the ImageScope software at the Dana-Farber/Harvard Cancer Center Tissue Microarray and Imaging Core. Tumor sections for immunofluorescent staining was performed as above, but with 4’,6-diamidino-2-phenylindole (DAPI, Life Technologies) and phycoerythrin (PE) mouse anti-human CD56 (MEM188, BioLegend). Fluorescent images were acquired on a Leica confocal microscope (TCS-SP5-AOBS) and images overlaid using Leica application suite advanced fluorescence software.
Statistical analysis
Data was analyzed using two-sided unpaired Student’s t-test or two-way ANOVA. *, **, and *** indicate p < 0.05, 0.01 and 0.005, respectively. All values and bars are represented as mean +/− standard deviation (S.D.) or standard error of the mean (S.E.M.).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
WAM, DKC, and RJM. contributed to the design of the experiments and drafted the manuscript. DKC. carried out experiments and analyzed data. RJM performed the cytotoxicity, animal study, and FACS staining and analyzed data. ZX performed the cytotoxicity, transwell, and migration studies. JS participated in animal study. SS conceived the study and participated histology staining. QZ conceived the study and helped draft the manuscript. All authors read and approved the final manuscript.
Funding
National Foundation for Cancer Research.