Recent progress in immunotherapy with immune checkpoint inhibitors has offered new options for the treatment of lung cancer, a deadly disease that each year causes approximately 155,870 deaths in the United States and 1.6 million deaths worldwide [
1,
2]. The anti-PD-1 antibodies nivolumab [
3,
4] and pembrolizumab [
5,
6] and the anti-PD-L1 antibodies atezolizumab [
7] and durvalumab [
8] have recently been approved for treatment of non-small cell lung cancer (NSCLC) by the US Food and Drug Administration. Clinical studies have shown that immune checkpoint blockade therapy for patients with advanced NSCLC improved survival rates, prolonged duration of response, and reduced treatment-related adverse effects [
9]. Nevertheless, despite the results of clinical trials with PD-1 [
4,
5,
10,
11] and PD-L1 [
7,
12,
13] inhibitors have been promising, the overall objective response rates in NSCLC patients who received such therapy were about 20%, and most patients had primary resistance [
3‐
5,
7,
10‐
12]. Strategies to overcome primary resistance to immune checkpoint inhibitors, either through combination therapy or through discovery of new targets to modulate anticancer immunity, are urgently needed so that more lung cancer patients can benefit from immunotherapy. However, preclinical investigation of therapeutic agents that recognize only the human version of immune checkpoint molecules is drastically limited by a lack of proper preclinical models due to interspecies differences in the specificity of checkpoint molecules, growth factors, and cytokines. Thus, mouse models with an in vivo “humanized” environment will be desirable for preclinical evaluation of strategies aimed at improving efficacies of immune-modulating drugs.
Patient-derived xenograft (PDX) models have been shown to recapitulate histologic features, gene expression patterns, and genomic alterations in human primary tumors [
14‐
16] and have emerged as robust preclinical models for drug development, molecular characterization of cancers, and strategic development of precision therapy [
17‐
21]. Our recent study on molecular characterization of lung cancer PDXs found that 93% of mutations found in primary tumors were also found in PDXs, suggesting that PDXs have the ability to recapitulate the mutations in the primary tumors [
22]. However, because PDXs are grown in immunodeficient mouse strains, they are regarded as inappropriate for preclinical evaluation of anticancer immunotherapy because of a lack of host immune components. Nevertheless, studies by others have shown that a subgroup of lung cancer patients has intensive lymphocyte infiltration in the tumor stroma and/or tumor nest [
23‐
26]. For example, a study with more than 1500 patients with resectable NSCLC found that about 10% of these patients had intense lymphocytic infiltration in their tumors, and this subset of patients had better overall survival outcomes than did patients with nonintense tumor lymphocytic infiltration [
25]. Evidence has shown that co-transplantation of human immune cells within metastatic melanomas into the subcutis of NOD-SCID IL-2Rγ
null (NSG) mice for generating PDXs may lead to graft-versus-host disease, reactive lymphoid infiltrates effacing xenografted tumors, and post-transplant B cell lymphomas associated with Epstein–Barr virus reactivation [
27]. It is also reported that intratumoral sustained release of recombinant human IL-12 in human lung tumor tissue implanted into NSG mice resulted in prolonged existence of effector memory T cells and CD138
+ plasma cells within the tumor xenograft (for up to 9 weeks) [
28]. The tumor-associated T cells were also found to migrate from the xenograft to the spleen, lung, and liver of the xenograft-bearing mice, and significant levels of human interferon-γ and immunoglobulin were detected in sera from these mice [
28], suggesting that human immune cells co-implanted with tumor tissues could survive and expand in immunodeficient mice. However, it remains undetermined whether patient-derived tumor-infiltrating lymphocytes (TILs) can co-exist within PDXs grown in NSG mice and how long these cells can persist.
To test whether patient-derived immune components co-exist in PDXs, we cultured PDX tumor tissues for TILs in vitro by adding human interleukin-2 (IL-2) to the medium. Our results showed that patient-derived TILs were successfully cultured from PDX samples. The mean time of PDX growth in NSG mice before harvesting to obtain TILs in vitro was about 113 days (range 63–292 days). Moreover, patient-derived TILs can be passed together with PDXs in early passages.