Studies using cell lines have been invaluable in providing information with regard to the molecular basis of drug resistance in most solid tumors. Studies in HNSCC cell lines have identified several classes of genes as paramount in drug resistance. Global transcript profiling of sensitive and resistant cell lines identified MMP-7 and MMP-13 as important for intrinsic cisplatin sensitivity (ICS) [9]. Correlation of marker profiles between the cisplatin-sensitive and resistant cell lines also indicated a deregulation of markers involved in DNA repair (RECQL, [10]; PMID: 25327479), cell cycle (CCND1, CCND3) [11] and miRNA profiles [12]. Studies in cell lines have implicated other pathways involved in EMT and metastasis (Twist-miR181-a; MACC1, YAP1), glucose transport (GLUT1) and CSC marker (CD147, CD44) [PMIDs: 25846049, 25501015, 24148247, 25421538, 25327479, 23617290, 23413783]. EGFR is another molecule that has been identified to influence the resistance to the TPF regimen in vitro; knockdown of the molecule in cell lines increases sensitivity to the drugs [13]. TPF-resistant cell lines are also known to overexpress markers of cell cycle regulation (survivin), DNA repair (ERCC1) and MDR (MDR1 and PgP-2) [14]. In vitro studies have also revealed the HNSCC cells with varied genetic background may need modified approaches; cells with mutant p53 can be sensitized to cisplatin by inhibition of checkpoint kinases such as Wee-1 [PMID: 25504633] (Tables 11.1 and 11.2).

Alterations in p53, one of the most prevalent in HNSCC, as per The Cancer Genome Atlas (TCGA), have been identified as prime in resistance to drug. Assessment of loss of heterozygosity (LOH) and microsatellite instability (MSI) in HNSCC patients, identified LOH at 9p or 17p (location of p53) to be significantly associated with drug resistance, LOH at 17p being predictive of low response to platinum-5-FU chemotherapy [15]. Mutations in exons 4–9 of p53 were of high prevalence in HNSCC (~68 %) with these patients showing no response to neoadjuvant therapy [16], mutations being more specifically predictive of response in patients with laryngeal [17] and maxillary squamous cell cancers [18]. Alterations in the codon 72 of p53, encoding either arginine (72R) or proline (72P), effect response to cisplatin; patients with 72R showing extremely low response rates [19].

Expression of p53 (at protein level) in combination with thymidylate synthase (TS) and GST-pi also significantly correlated with response and survival to cisplatin-based neo-adjuvant chemotherapy NACT [20]. Immunohistochemical levels of GSTs when scored against chemotherapy response of patients showed that overall response rate (CR + PR) in patients with low GST was 88 %, while those with high GST scores only showed 19 % response (p = 0.0001). Among the subset treated with NACT, 100 % response was observed in GST low patients, while in patients treated with chemotherapy for relapsed disease, the response rate was 70 % [21]. When the genes involved in thymidine synthesis and metabolism were assessed for their association with response to combination therapy (radiation with induction CT of platinum/5-FU or concurrent platinum therapy), it was observed that lesser percentage of cells with nuclear thymidine phosphorylase (TP) in pretreatment biopsies was associated with high rate of complete response (CR). These patients also showed a high relapse-free survival (p = 0.001) [22].

Multidrug resistance proteins, MDR1 and MRP, are associated with resistance to therapy in oral cancer; expression of these proteins was high in the nonresponders as compared to responders in patients with tongue cancer [23]. Another study reported the induction of PgP by radiation and may explain the low response rates of patients who are subjected to concurrent chemotherapy [24]. Assessment of patients with advanced cancers of the head and neck also showed that nonresponders had a high expression of PgP and MDR-3 [25].

Markers of DNA repair are extensively correlated to treatment response; studies associating response to molecular profile in patients showed that increased co-expression of SNAIL and ERCC1 (SNAIL is a known activator of ERCC1) contributes toward resistance to cisplatin and was also found to correlate with prognosis in patients with HNSCC [26]. Expression of ERCC1 along with Bcl-2 and other multidrug-resistant genes (MDR-1, MRP-1 and ATP7B) were downregulated in recurrent maxillary disease [18]. Analysis of ERCC1 alterations at mutation, protein and transcript levels indicated an increased association with treatment outcome. In patients treated with adjuvant cisplatin-based chemoradiation, high IHC scores (H-score) of ERCC1 protein correlated with response and survival [27]. Further, in patients treated with cisplatin-based induction therapy followed by concurrent chemoradiation, low immunohistochemical levels of ERCC1 correlated with increased 12-month disease-free (73.3 % vs. 42.3 %, p < 0.001) and overall survival [28]. Other members of the ERCC family, XPF, or ERCC4 levels in IHC are also shown to correlate with progression-free survival in patients treated with DNA-damaging agents [29].

Deregulation of cell cycle and apoptotic proteins plays a major role in carcinogenesis as well as resistance to drugs. In patients exposed to neoadjuvant therapy (cisplatin based), low CCND1 expression indicated better response (p < 0.0001) with significantly better OS and DFS (73 % vs. 8 %, p < 0.001; 63 % vs. 6 %, p < 0.001); there was no correlation in patients treated with surgery and radiation indicating that this marker might be predictive in selecting patients for NACT [30]. Tumors treated with cisplatin, positive for the apoptotic protein, Bcl-2, prior to treatment had a high risk of treatment failure with high hazard ratio (hazard ratio, 5.99; 95 % confidence interval, 1.73–20.8; P = 0.0014) [31]. Studies have also shown increased expression of the Bcl-XL to correlate with low DFS in patients with advanced disease [25]. X-linked inhibitor of apoptosis (XIAP) was also associated with resistance to cisplatin (p = 0.036) and outcome. Additionally, it was observed that the expression of this gene was further induced in patients exposed to the therapy, which lead to a poorer outcome in this subset [32].

Alterations in the targets of the cytotoxic agents are a major mechanism of drug resistance adopted, especially in cases of taxanes. Increased expression of acetylated tubulin is known to correlate with occurrence of lymph node metastasis in HNSCC patients treated with docetaxel along with cisplatin and 5-FU. Extensive evidence correlating tubulin modifications with taxane resistance is available in other cancers; β-tubulin mutations in serum DNA were associated with paclitaxel resistance in non-small cell lung cancer (NSCLC) [33].

Many other markers have shown correlation with treatment response; polymorphisms in matrix metalloproteinase 3 (1612insA) promoter; 6A/6A genotype (wherein both the alleles have 6 adenosines at the position due to the insertion) was an independent response factor for patients treated with 5-FU-cisplatin therapy as compared to 5A/6A (heterozygous with one allele having no insertion) or 5A/5A genotypes (wild type allele with no insertions at the position) [34]. Other studies in patients, treated with 5-FU or cisplatin and paclitaxel, have shown increased expression of c-erbB2 to correlate with progression-free and overall survival [35]. Among other markers, downregulation of miR34a, involved in regulation of silent information regulator 1 (SIRT1) and thereby p53, was associated with poor DFS and local control rates in HNSCC patients treated with cisplatin [36] (Fig. 11.3).

The TPF combination therapy regime is considered as the gold standard of induction chemotherapy in head and neck cancer, but resistance to this regime has also been a major challenge. In patients with clinically positive nodes (clinically node positive, cN+), high cyclin D1 expression was shown to predict benefit from addition of TPF regimen to standard treatment [37]. Expression of growth-dependent factor (GDF15) is shown to correlate with poor prognosis in patients treated with TPF therapy [38]. Low GDF predicts good overall as well as distant metastasis-free survival in these patients. Other studies in TPF-treated patients, evaluating marker correlation with response, indicated that acetylated tubulin (AT), annexin (low) and CDK1 (high) were predictors of response [39, 40]. In advanced recurrent patients treated with taxol and platinum, low levels of ERCC1 and increased RASSF1A were predictive of good prognosis [41]. Molecular profiling of oropharyngeal patients in the TAX324 trial (comparing TPF against PF in locally advanced HNSCC) showed that low beta tubulin II (beta T-2) was predictive of therapy benefit [42]. Categorization of the oropharyngeal patients of this trial based on their HPV status indicated that the presence of a specific molecular profile (high beta T-II, GST-pi and p53 along with low Bcl-2) was predictive of survival rates when treated with TPF/PF [43]. This observation emphasized the significance of risk stratification prior to treatment selection. A concept further gaining significance is the role of cancer stem cells in imparting drug resistance in cancers including head and neck cancer; studies have shown that in patients treated with the TPF regimen of drugs, expression levels of CD44, BMI1 and Notch1 correlated with recurrence. In resistant patients, a consequent enrichment of these markers posttreatment is a further indicator of poor response [44]. This aspect of the CSC-mediated drug resistance has been discussed in detail in the chapter 14.

EGFR, 170 KDa cell surface receptor protein with extracellular ligand binding and an intracellular domain with tyrosine kinase activity [47], aids in the normal development and proliferation of epithelial tissue and is found to be overexpressed in 80–90 % of head and neck cancer. EGFR overexpression, correlated with reduced survival, poor prognosis, and aggressive disease [52], is reported to be responsible for cell proliferation, differentiation, antiapoptotic signaling, angiogenesis and metastasis in various malignancies. EGFR is activated through ligand binding which leads to autophosphorylation of the tyrosine kinase domain. The various EGFR ligands are TGF-α, amphiregulin, epiregulin, EGF, epigen, betacellulin and heparin-binding EGFR. The EGFR pathway has different signaling routes: P13K-PDK1-AKT, Ras-Raf-MEK-ERK (mediates transcription and activates other proteins), PLCγ-PKC (cell cycle progression), and JAK-STAT (cell proliferation, survival, transformation) [53, 54] (Fig. 11.4).

Targeting EGFR pathway, in patients wherein it is dysregulated, is hence considered a viable therapeutic option due to its increased prevalence in the patients as well it is functional relevance. The approaches towards EGFR targeting includes i) blockage of ligand binding using monoclonal antibodies (Cetuximab and Panitumumab) ii) blockage of its TKI domain (Geftinib, Erlotinib) activity thereby blocking the downstream signaling molecules. Nevertheless, the response rates in patients to these various therapies ranges between 10–40%, either due to primary or acquired resistance. The possible resistance mechanisms for EGFR therapy include mutation of EGFR, oncogenic shift or activation of a bypass pathway, and/or due to modification of any pathway/molecule essential for EGFR TKI-mediated apoptosis [55] (Fig. 11.5).

Targeting the EGFR protein is primarily through the use of monoclonal antibodies; Cetuximab and panitumumab are used for blocking the ligand binding to the receptor; they bind with a higher affinity as compared with the regular ligands. Cetuximab binds specifically to the extracellular domain [56]. Correlation of the marker profile with response to the drug indicated that expression of EFGR protein was indicative of response in the trials such as E2303 [Eastern Cooperative Oncology Group (ECOG)], Erbitux in First-Line Treatment of Recurrent or Metastatic Head and Neck Cancer (EXTREME), and Cetuximab combined with Irinotecan in first line therapy for metastatic colorectal cancer(CRYSTAL) carried out in patients [57, 58]. Further assessment of the patients identified extracellular ERK1/2 and activated RAS/MAPK/ERK and/or PI3K/AKT pathways as indicative of PFS and OS [59]. However, marker evaluation in EXTREME trial patients showed no association of EGFR copy number with response in metastatic patients [60, 61]. Studies also revealed that serum biomarkers (VEGF, IL-6) were associated with response to cetuximab combination therapy with platinum/taxol) [62]. KRAS mutation (p.Gly12Val) and somatic EGFR mutation located in exon 19 is also suggested to contribute to the limited clinical response to cetuximab therapy in combination with radiotherapy; in the same study, high EGFR expression was indicative of increased treatment response [63]. Additionally, mutation in the 3′ UTR of KRAS have been associated with treatment response; patients (recurrent/metastatic HNSCC with mutation in 3′ UTR KRAS) showed improved response when treated with cetuximab and platinum as compared to cisplatin treatment alone [64].